| *** spandhe has quit IRC | 00:05 | |
| *** jaybeale has joined #openstack-ironic | 00:07 | |
| *** garthb has quit IRC | 00:08 | |
| *** blakec has quit IRC | 00:08 | |
| *** cdearborn has quit IRC | 00:12 | |
| *** [2]cdearborn is now known as cdearborn | 00:12 | |
| *** rama_y has quit IRC | 00:16 | |
| jlvillal | jroll: Done: https://bugs.launchpad.net/ironic/+bug/1560264 | 00:16 |
|---|---|---|
| openstack | Launchpad bug 1560264 in Ironic "spawn_n() should have request context set" [Undecided,New] | 00:16 |
| jlvillal | lintan: ^^^ | 00:17 |
| *** daemontool_ has joined #openstack-ironic | 00:26 | |
| *** daemontool has quit IRC | 00:27 | |
| *** izaakk has quit IRC | 00:27 | |
| *** praneshp_ has quit IRC | 00:30 | |
| jroll | jlvillal: thanks | 00:34 |
| *** aginwala has joined #openstack-ironic | 00:35 | |
| lintan | morning jivilla | 00:46 |
| *** cdearborn has quit IRC | 00:48 | |
| lintan | thanks jivillal to create a bug for that | 00:48 |
| *** jaybeale has quit IRC | 00:54 | |
| *** hoangcx has joined #openstack-ironic | 00:54 | |
| openstackgerrit | Naohiro Tamura proposed openstack/ironic: Fix NamedTemporaryFile() OSError Exception https://review.openstack.org/293762 | 00:55 |
| lintan | jroll: I will make it today :) | 00:55 |
| *** dims has joined #openstack-ironic | 00:57 | |
| *** Sukhdev has quit IRC | 01:09 | |
| *** mtanino has quit IRC | 01:10 | |
| *** thrash is now known as thrash|pt0 | 01:13 | |
| dims | jroll : still around? | 01:30 |
| dims | jroll : https://review.openstack.org/#/c/295559/ | 01:31 |
| jroll | dims: sup | 01:32 |
| dims | jroll : 1.2.0 is to be stable/mitaka right? | 01:32 |
| jroll | dims: yep, hence the commit message :) | 01:32 |
| dims | jroll : yep :) | 01:33 |
| dims | jroll : list-changes output looks good? right - http://logs.openstack.org/59/295559/1/check/gate-releases-tox-list-changes/af057ff/console.html#_2016-03-21_23_01_09_360 | 01:33 |
| dims | just sanity checks :) | 01:33 |
| jroll | dims: yep, looks as expected | 01:34 |
| dims | thanks. i'll cut the release now | 01:34 |
| jroll | <3 | 01:34 |
| jroll | dims: fyi, ironic, ironic-ui, bifrost releases all coming this week, shooting for wednesday-ish | 01:34 |
| jroll | all stable/mitaka | 01:34 |
| dims | jroll : ack | 01:35 |
| *** aginwala has quit IRC | 01:35 | |
| *** aginwala has joined #openstack-ironic | 01:35 | |
| *** ChrisAusten has joined #openstack-ironic | 01:37 | |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/ironic-python-agent: Update reno for stable/mitaka https://review.openstack.org/295591 | 01:38 |
| *** Nisha has joined #openstack-ironic | 01:42 | |
| dims | jroll : feel free to edit that or create a new review - release scripts generate that so folks know reno needs to be setup | 01:45 |
| dims | jroll : this may be a simpler one https://review.openstack.org/#/c/295590/ | 01:45 |
| jroll | dims: yeah, I'll do that, approved that other one. thanks. | 01:46 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic-python-agent: Update reno for stable/mitaka https://review.openstack.org/295591 | 01:48 |
| *** vishwanathj has quit IRC | 01:50 | |
| *** spandhe has joined #openstack-ironic | 01:53 | |
| *** lucas-dinner has quit IRC | 01:55 | |
| *** Marga_ has quit IRC | 01:55 | |
| *** Marga_ has joined #openstack-ironic | 01:57 | |
| *** lucasagomes has joined #openstack-ironic | 02:00 | |
| *** Marga_ has quit IRC | 02:01 | |
| *** aginwala has quit IRC | 02:02 | |
| *** baoli has quit IRC | 02:03 | |
| *** baoli has joined #openstack-ironic | 02:04 | |
| *** vishwanathj has joined #openstack-ironic | 02:05 | |
| *** baoli has quit IRC | 02:06 | |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic-python-agent: Update reno for stable/mitaka https://review.openstack.org/295591 | 02:09 |
| *** jaybeale has joined #openstack-ironic | 02:10 | |
| *** jaybeale has quit IRC | 02:15 | |
| *** aginwala has joined #openstack-ironic | 02:15 | |
| *** baoli has joined #openstack-ironic | 02:17 | |
| *** dims has quit IRC | 02:28 | |
| *** Marga_ has joined #openstack-ironic | 02:28 | |
| *** Marga_ has quit IRC | 02:33 | |
| *** baoli_ has joined #openstack-ironic | 02:33 | |
| *** baoli has quit IRC | 02:37 | |
| *** rbudden has joined #openstack-ironic | 02:42 | |
| *** Nisha has quit IRC | 02:42 | |
| *** Fdaisuke has joined #openstack-ironic | 02:43 | |
| *** aginwala has quit IRC | 02:44 | |
| *** vishwana_ has joined #openstack-ironic | 02:47 | |
| *** vishwana_ is now known as vishwanathj__ | 02:48 | |
| *** vishwanathj has quit IRC | 02:50 | |
| *** vishwanathj__ is now known as vishwanathj | 02:53 | |
| *** lazy_prince has joined #openstack-ironic | 02:58 | |
| *** hoangcx has quit IRC | 02:59 | |
| *** hoangcx has joined #openstack-ironic | 02:59 | |
| *** harshs has joined #openstack-ironic | 03:03 | |
| *** baoli_ has quit IRC | 03:05 | |
| *** baoli has joined #openstack-ironic | 03:06 | |
| *** Marga_ has joined #openstack-ironic | 03:10 | |
| *** vishwanathj is now known as vishwanathj_zzz | 03:10 | |
| openstackgerrit | Tan Lin proposed openstack/ironic: Append 'Openstack-Baremetal-Request-ID' header to the response https://review.openstack.org/238008 | 03:11 |
| *** harshs_ has joined #openstack-ironic | 03:13 | |
| *** Marga_ has quit IRC | 03:14 | |
| *** harshs has quit IRC | 03:15 | |
| *** harshs_ is now known as harshs | 03:15 | |
| *** rbudden has quit IRC | 03:15 | |
| *** baoli has quit IRC | 03:17 | |
| *** baoli has joined #openstack-ironic | 03:17 | |
| lazy_prince | lucasagomes: did you get a chance to review https://review.openstack.org/#/c/287784 | 03:18 |
| *** aginwala has joined #openstack-ironic | 03:19 | |
| *** links has joined #openstack-ironic | 03:19 | |
| *** phuongnh has joined #openstack-ironic | 03:21 | |
| *** baoli has quit IRC | 03:22 | |
| *** aginwala has quit IRC | 03:23 | |
| *** baoli has joined #openstack-ironic | 03:23 | |
| *** yuanying has quit IRC | 03:28 | |
| *** Marga_ has joined #openstack-ironic | 03:31 | |
| *** praneshp_ has joined #openstack-ironic | 03:33 | |
| *** Marga_ has quit IRC | 03:35 | |
| *** yuanying has joined #openstack-ironic | 03:38 | |
| *** achanda has quit IRC | 03:43 | |
| *** Sukhdev has joined #openstack-ironic | 03:47 | |
| *** Marga_ has joined #openstack-ironic | 03:52 | |
| *** Marga_ has quit IRC | 04:01 | |
| *** aswadr_ has joined #openstack-ironic | 04:01 | |
| *** Marga_ has joined #openstack-ironic | 04:01 | |
| *** baoli has quit IRC | 04:03 | |
| *** yuanying has quit IRC | 04:04 | |
| *** yuanying has joined #openstack-ironic | 04:06 | |
| *** chlong|wfh has quit IRC | 04:12 | |
| *** Sukhdev has quit IRC | 04:21 | |
| *** Sukhdev has joined #openstack-ironic | 04:21 | |
| *** lazy_prince has quit IRC | 04:21 | |
| *** chlong has joined #openstack-ironic | 04:25 | |
| *** achanda has joined #openstack-ironic | 04:32 | |
| *** suro-patz has joined #openstack-ironic | 04:35 | |
| *** saripurigopi has joined #openstack-ironic | 04:46 | |
| *** spandhe has quit IRC | 04:49 | |
| *** daemontool_ has quit IRC | 04:53 | |
| *** daemontool__ has joined #openstack-ironic | 04:53 | |
| openstackgerrit | Shivanand Tendulker proposed openstack/ironic: Support configdrive in iscsi deploy for raw images https://review.openstack.org/225115 | 04:57 |
| *** harshs has quit IRC | 05:05 | |
| *** harshs has joined #openstack-ironic | 05:07 | |
| *** suro-patz has quit IRC | 05:18 | |
| openstackgerrit | Shivanand Tendulker proposed openstack/ironic-lib: Support configdrive in iscsi deploy for raw images https://review.openstack.org/230924 | 05:18 |
| *** suro-patz has joined #openstack-ironic | 05:20 | |
| *** ChrisAusten has quit IRC | 05:25 | |
| *** suro-patz has quit IRC | 05:26 | |
| *** jaybeale has joined #openstack-ironic | 05:28 | |
| *** saripurigopi has quit IRC | 05:41 | |
| openstackgerrit | Sivaramakrishna Garimella proposed openstack/ironic: WIP: add network drivers https://review.openstack.org/285852 | 05:42 |
| *** praneshp_ has quit IRC | 05:46 | |
| *** praneshp_ has joined #openstack-ironic | 05:46 | |
| *** suro-patz has joined #openstack-ironic | 05:48 | |
| *** harlowja_at_home has quit IRC | 05:52 | |
| *** achanda has quit IRC | 05:55 | |
| *** lintan has quit IRC | 06:01 | |
| *** aginwala has joined #openstack-ironic | 06:03 | |
| *** lintan has joined #openstack-ironic | 06:06 | |
| *** ohamada has joined #openstack-ironic | 06:11 | |
| *** hoangcx has quit IRC | 06:15 | |
| *** hoangcx has joined #openstack-ironic | 06:16 | |
| *** ohamada has quit IRC | 06:25 | |
| *** ChubYann has quit IRC | 06:25 | |
| *** jaybeale has quit IRC | 06:30 | |
| *** Haomeng has quit IRC | 06:32 | |
| *** praneshp_ has quit IRC | 06:46 | |
| *** praneshp__ has joined #openstack-ironic | 06:46 | |
| *** Fdaisuke_ has joined #openstack-ironic | 06:54 | |
| *** Fdaisuke has quit IRC | 06:55 | |
| *** hoangcx has quit IRC | 06:56 | |
| *** hoangcx has joined #openstack-ironic | 06:56 | |
| *** davidlenwell has quit IRC | 07:01 | |
| *** praneshp__ has quit IRC | 07:03 | |
| *** aginwala has quit IRC | 07:07 | |
| *** davidlenwell has joined #openstack-ironic | 07:09 | |
| *** Sukhdev has quit IRC | 07:11 | |
| *** mkovacik__ has quit IRC | 07:21 | |
| *** ohamada has joined #openstack-ironic | 07:22 | |
| *** Haomeng has joined #openstack-ironic | 07:26 | |
| *** tesseract has joined #openstack-ironic | 07:28 | |
| *** tesseract is now known as Guest91251 | 07:28 | |
| *** hoangcx_ has joined #openstack-ironic | 07:30 | |
| *** hoangcx has quit IRC | 07:31 | |
| *** hoangcx_ is now known as hoangcx | 07:32 | |
| *** rcernin has joined #openstack-ironic | 07:39 | |
| *** jtomasek has joined #openstack-ironic | 07:39 | |
| *** moshele has joined #openstack-ironic | 07:49 | |
| *** daemontool__ has quit IRC | 07:50 | |
| openstackgerrit | Tan Lin proposed openstack/ironic: Adopt Ironic's own context https://review.openstack.org/295663 | 08:18 |
| *** suro-patz has quit IRC | 08:19 | |
| *** athomas has joined #openstack-ironic | 08:23 | |
| *** ifarkas has joined #openstack-ironic | 08:27 | |
| *** openstackgerrit has quit IRC | 08:33 | |
| *** openstackgerrit has joined #openstack-ironic | 08:34 | |
| *** daemontool has joined #openstack-ironic | 08:34 | |
| openstackgerrit | Naohiro Tamura proposed openstack/ironic: iRMC power driver for soft power off and inject nmi https://review.openstack.org/216743 | 08:35 |
| *** Fdaisuke_ has quit IRC | 08:40 | |
| *** pcaruana has joined #openstack-ironic | 08:40 | |
| *** _degorenko|afk is now known as degorenko | 08:41 | |
| *** harshs has quit IRC | 08:46 | |
| phuongnh | Hi all, while delploying Ironic via devstack, I got the error msg "server didn't become ssh-able!", devstack stops at line 848 Ironic file, please guide me to fix it | 08:46 |
| *** lazy_prince has joined #openstack-ironic | 08:51 | |
| *** daemontool has quit IRC | 08:57 | |
| *** achanda has joined #openstack-ironic | 09:02 | |
| *** mbound has joined #openstack-ironic | 09:02 | |
| *** achanda has quit IRC | 09:02 | |
| *** mgoddard has joined #openstack-ironic | 09:03 | |
| openstackgerrit | Zhenguo Niu proposed openstack/ironic: Restart consoles on conductor startup https://review.openstack.org/257252 | 09:05 |
| lucasagomes | lazy_prince, hi there, not yesterday, will do now | 09:13 |
| *** hoangcx has quit IRC | 09:13 | |
| openstackgerrit | lokesh s proposed openstack/ironic: Add support for the audit middleware https://review.openstack.org/272658 | 09:15 |
| *** derekh has joined #openstack-ironic | 09:17 | |
| *** hoangcx has joined #openstack-ironic | 09:18 | |
| *** ndipanov has quit IRC | 09:18 | |
| *** ndipanov has joined #openstack-ironic | 09:18 | |
| *** ndipanov has quit IRC | 09:18 | |
| *** ndipanov has joined #openstack-ironic | 09:19 | |
| *** Nisha has joined #openstack-ironic | 09:21 | |
| *** e0ne has joined #openstack-ironic | 09:22 | |
| *** vinm213 has joined #openstack-ironic | 09:25 | |
| vinm213 | morning ironicers | 09:25 |
| pas-ha | morning Ironic | 09:27 |
| lucasagomes | morning | 09:28 |
| * lucasagomes is sad for brussels :-/ | 09:28 | |
| Nisha | morning Ironic!!! | 09:29 |
| ifarkas | morning all | 09:30 |
| ifarkas | lucasagomes, yeah, that's very sad :-( | 09:30 |
| * vinm213 too sad about the incident | 09:31 | |
| *** mkovacik__ has joined #openstack-ironic | 09:32 | |
| *** div has joined #openstack-ironic | 09:33 | |
| div | i am facing issue with devstack install | 09:33 |
| div | what is the local.conf parameter for install ironic in hardware nodes | 09:35 |
| div | can anybody tell | 09:35 |
| vinm213 | enable_plugin ironic https://git.openstack.org/openstack/ironic | 09:36 |
| vinm213 | IRONIC_USING_PLUGIN=true | 09:36 |
| *** athomas has quit IRC | 09:36 | |
| vinm213 | enable_service ironic ir-api ir-cond | 09:36 |
| vinm213 | i think u need these things in local.conf | 09:36 |
| pas-ha | you do not need IRONIC_USE_PLUGIN any more | 09:37 |
| div | enable_service ironic enable_service ir-api enable_service ir-cond | 09:38 |
| div | i have enabled these | 09:38 |
| pas-ha | div: IRONIC_IS_HARDWARE=True is probably what you need | 09:38 |
| div | is there anything specif for baremetal nodes need to be added | 09:38 |
| alineb | morning all | 09:39 |
| div | thanks pas-ha, let me try with it | 09:40 |
| *** athomas has joined #openstack-ironic | 09:41 | |
| div | and the error i am getting during devstack install is, | 09:41 |
| div | openstack object store account set --property Temp-URL-Key=password 2016-03-21 09:36:49.425 | Not Found (HTTP 404) 2016-03-21 09:36:49.453 | + /home/stack/devstack/lib/swift:swift_configure_tempurls:L1: exit_trap 2016-03-21 09:36:49.453 | + ./stack.sh:exit_trap:L474: local r=1 2016-03-21 09:36:49.454 | ++ ./stack.sh:exit_trap:L475: jobs -p 2016-03-21 09:36:49.454 | + ./stack.sh:exit_trap:L475: jobs= 2016-03-21 09:36:49 | 09:41 |
| div | is swift missing any configuration in local.conf, so that it is failing with 404? | 09:42 |
| *** electrofelix has joined #openstack-ironic | 09:44 | |
| vinm213 | are u running latest devstack and updated your requirements folder. | 09:45 |
| div | i have cloned the latest devstack | 09:46 |
| div | can you tell me how to update the requirement since i am new to devstack | 09:46 |
| vinm213 | if u had already run devstack once u see some directories in /opt/stack | 09:48 |
| pas-ha | swift is not installed by default in devstack afaik, you have to enable it manually | 09:48 |
| pas-ha | div: just for inspiration you might take a look at my sample for local.conf, uncomment stuff you need https://github.com/pshchelo/stackdev/blob/master/local.conf.sample | 09:49 |
| pas-ha | generally Ironic requires Keystone, Neutron, Nova, Glance and Swift | 09:50 |
| *** sambetts|afk is now known as sambetts | 09:50 | |
| vinm213 | but if it was not enabled, i wonder should we see that error? I think in your case it is enabled | 09:50 |
| *** dims has joined #openstack-ironic | 09:51 | |
| pas-ha | vinm213: not reallly. it seems swift temp url key has a default | 09:51 |
| div | i have enabled swift | 09:52 |
| div | enable_service s-proxy enable_service s-object enable_service s-container enable_service s-account | 09:52 |
| pas-ha | also need SWIFT_ENABLE_TEMPURLS=True | 09:53 |
| div | SWIFT_HASH=password SWIFT_TEMPURL_KEY=password | 09:53 |
| div | yes..SWIFT_ENABLE_TEMPURLS=True | 09:53 |
| div | i have set this as well | 09:53 |
| div | but still i fails | 09:53 |
| pas-ha | hmm.. then I'm out of ideas :( | 09:53 |
| phuongnh | Dear all, while deploying Ironic via devstack, I got the error msg "server didn't become ssh-able!", devstack stops at line 848 stack/ironic/devstack/lib/ironic file and I cannot continue | 09:57 |
| sambetts | Morning all | 09:57 |
| phuongnh | my deployment model is 2 VMs (Master VM and Guest VM) on KVM, Ironic is installed on the Master VM and try to deploy on Guest VM | 09:59 |
| phuongnh | I use the same local.conf file as on deployment document, just change password and network's configuration | 10:00 |
| phuongnh | but I always fail at that line of code | 10:01 |
| div | : Unable to locate package liberasurecode-dev | 10:01 |
| div | devstack fails with this error | 10:01 |
| div | can anybody help | 10:01 |
| openstackgerrit | Sivaramakrishna Garimella proposed openstack/ironic: Add portgroups to support LAG interfaces - API https://review.openstack.org/206244 | 10:03 |
| openstackgerrit | Sivaramakrishna Garimella proposed openstack/ironic: Update the deploy drivers with network flipping logic https://review.openstack.org/213262 | 10:03 |
| openstackgerrit | Sivaramakrishna Garimella proposed openstack/ironic: WIP: add network drivers https://review.openstack.org/285852 | 10:03 |
| *** dims has quit IRC | 10:04 | |
| *** mgould has joined #openstack-ironic | 10:04 | |
| lazy_prince | lucasagomes: thanks for review.. | 10:10 |
| lucasagomes | lazy_prince, no problem, I didn't test it tho, but looks sane | 10:10 |
| mkovacik__ | morning Ironic! | 10:11 |
| lazy_prince | now only if we can just get attention from tripleo guys to get it merged.. | 10:11 |
| lucasagomes | lazy_prince, o/ | 10:11 |
| lucasagomes | lazy_prince, one thing, actually I forgot to ask the author. There's an element called "grub2" that does some of that | 10:12 |
| lucasagomes | at least when it comes to package installation | 10:12 |
| lucasagomes | maybe we should just merge the bootloader element with that grub2 element | 10:12 |
| *** dims has joined #openstack-ironic | 10:12 | |
| vinm213 | div: you need to get the packages from source on to you'r system and install them using "dpkg -i ..." | 10:13 |
| openstackgerrit | Milan Kováčik proposed openstack/ironic-inspector-specs: High Availability for Ironic Inspector https://review.openstack.org/253675 | 10:15 |
| sambetts | phuongnh: Do you have an ssh server running on the machine that your running devstack on ? | 10:16 |
| vinm213 | i also had similar issues earlier | 10:17 |
| sambetts | lucasagomes: I've been trying to get them to merge the tripleo localboot element and the grub2 element for a while but no one listened :( | 10:18 |
| lucasagomes | sambetts, ouch, yeah there's def an overlap there | 10:19 |
| lucasagomes | sambetts, did they give any reason not to merge it? | 10:19 |
| sambetts | lucasagomes: nope... I started pushing them to do it when the grub2 element got broken, and the localboot element already had the fix for it, but then they just fixed it in a different way in the grub2 element :/ | 10:20 |
| lucasagomes | ouch :-/ | 10:20 |
| lucasagomes | sambetts, I left a comment on the patch let's see if someone says something | 10:21 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Update authorized_keys with new key only https://review.openstack.org/295293 | 10:21 |
| sambetts | lucasagomes: link? | 10:21 |
| mkovacik__ | guys, I'd like to appeal on reviewing the HA for inspector spec again ;) https://review.openstack.org/#/c/253675 | 10:21 |
| sambetts | lucasagomes: found it in the scroll back :) | 10:22 |
| phuongnh | sambetts: yes, I can use PuTTY to connect to the machine normaly | 10:22 |
| lucasagomes | heh you were quicker! | 10:22 |
| sambetts | phuongnh: what driver are you planning on running with your devstack install? | 10:23 |
| mkovacik__ | lucasagomes, ifarkas, sambetts, aarefiev, devananda, mgould : https://review.openstack.org/#/c/253675 please ;) | 10:23 |
| vinm213 | i have a quick question, | 10:23 |
| sambetts | phuongnh: By default Ironic devstack will run the pxe_ssh driver to control VMs created on the devstack host | 10:23 |
| *** ipukha has joined #openstack-ironic | 10:24 | |
| phuongnh | sambetts: I use default setting on local.conf file. I use pxe_ssh driver in row IRONIC_DEPLOY_DRIVER=pxe_ssh | 10:24 |
| lazy_prince | lucasagomes: I will ask the author to look into grub2 element... | 10:26 |
| vinm213 | Do we require to validate the size of block device, https://github.com/openstack/ironic-python-agent/blob/master/ironic_python_agent/hardware.py#L489-L490 here | 10:26 |
| phuongnh | sambetts: the default local.conf file mean the file in this page: http://docs.openstack.org/developer/ironic/dev/dev-quickstart.html#deploying-ironic-with-devstack | 10:27 |
| vinm213 | i have a case where the /dev/sda size returned from "lsblk -Pbid" is returning lesser than this size | 10:27 |
| vinm213 | and the deployment of the node is failing with image download error | 10:27 |
| openstackgerrit | Nisha Agarwal proposed openstack/ironic: Document partition image support with agent_ilo https://review.openstack.org/295710 | 10:28 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Update authorized_keys with new key only https://review.openstack.org/295293 | 10:29 |
| lucasagomes | lazy_prince, thanks! | 10:29 |
| sambetts | lucasagomes: https://review.openstack.org/#/c/219612/ this was the patch that I discussed the localboot element merging with the grub2 one | 10:29 |
| lucasagomes | mkovacik__, ++ | 10:29 |
| *** achanda has joined #openstack-ironic | 10:31 | |
| *** achanda has quit IRC | 10:31 | |
| mkovacik__ | lucasagomes, sorry, what was it about? ;) | 10:31 |
| lucasagomes | mkoderer__, the spec for inspector | 10:31 |
| * lucasagomes is reading | 10:31 | |
| mkovacik__ | lucasagomes, :) cool, thx | 10:33 |
| *** sivaramakrishna has joined #openstack-ironic | 10:33 | |
| vinm213 | any one,any idea on this? | 10:34 |
| sambetts | phuongnh: can you please check that /opt/stack/data/ironic/ssh_keys/ironic_key and ironic_key.pub are being created successfully | 10:34 |
| phuongnh | sambetts: yes, they are created, but the ironic_key's right is rw------- and ironic_key.pub's right is rw-r--r--, is that correct? | 10:38 |
| *** sivaramakrishna has quit IRC | 10:40 | |
| sambetts | phuongnh: yes, thats all fine | 10:48 |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic: Agent: Out-of-band power off on deploy https://review.openstack.org/291829 | 10:49 |
| sambetts | phuongnh: can you now check if your .ssh/authorized_keys contains a copy of the ironic_key.pub file | 10:50 |
| *** hoangcx has quit IRC | 10:54 | |
| phuongnh | sambetts: I have remark the check and run devstack again. now 2 ironic_key files are gone and in .ssh folder I found authorized_keys newly created | 11:02 |
| Nisha | lucasagomes, hi | 11:04 |
| phuongnh | sambetts: I have to leave now, see you in next 2 hours | 11:04 |
| Nisha | lucasagomes, i was adding disk_label for partition image support for agent drivers...i have some query | 11:05 |
| lucasagomes | Nisha, hi there | 11:05 |
| Nisha | lucasagomes, suppose the user adds the boot mode as uefi but the disk_label as msdos, what is the result | 11:05 |
| TheJulia | Good morning everyone o/ | 11:06 |
| lucasagomes | Nisha, the disk will be partitioned with an MBR | 11:06 |
| lucasagomes | and will boot in uefi | 11:06 |
| Nisha | lucasagomes, will that work for paryition images as well? | 11:07 |
| Nisha | and if viceversa combination is given? | 11:07 |
| lucasagomes | Nisha, it should yes. UEFI does work with MBR partitions | 11:08 |
| lucasagomes | tho some OSs may have problems with that (Windows e.g) | 11:08 |
| lucasagomes | Nisha, http://superuser.com/questions/739153/uefi-with-mbr-partition-table?answertab=votes#tab-top | 11:08 |
| lucasagomes | or http://www.uefi.org/sites/default/files/resources/UEFI%202_5.pdf#G9.1345547 | 11:08 |
| Nisha | and if boot_mode as bios and disk_label as gpt? | 11:09 |
| Nisha | uefi with MBR is still valid combination but the other way its not | 11:09 |
| lucasagomes | Nisha, that's fine too, GPT has an MBR in it and is backwards compat | 11:09 |
| lucasagomes | tho, ofc there's always edge cases (some vendor BIOS may complain about it) | 11:09 |
| Nisha | lucasagomes, ohk. | 11:10 |
| Nisha | lucasagomes, thanks got it. | 11:10 |
| *** phuongnh has quit IRC | 11:10 | |
| lucasagomes | Nisha, for if you have a disk that is > 2 TB you may want to use BIOS + GPT | 11:10 |
| lucasagomes | Nisha, btw http://docs.openstack.org/developer/ironic/deploy/install-guide.html?highlight=disk_label#choosing-the-disk-label | 11:12 |
| Nisha | lucasagomes, ok. so the same combination shud be fine for "partition image support for agent drivers" | 11:12 |
| *** daemontool has joined #openstack-ironic | 11:13 | |
| lucasagomes | Nisha, yup | 11:14 |
| lucasagomes | the feature was added to the ironic-lib, so it covers both | 11:14 |
| lucasagomes | agent and iscsi | 11:14 |
| Nisha | yes , but work_on_disk() which is called in ironci for iscsi drivers has disk_label property, while for agent drivers work_on_disk() is called in IPA | 11:15 |
| lucasagomes | Nisha, oh, and IPA is not passing the disk_label parameter there? | 11:16 |
| * lucasagomes looks at the code | 11:16 | |
| Nisha | lucasagomes, no. As i wasnt sure of the testing part i didnt added it | 11:17 |
| *** Keedya has joined #openstack-ironic | 11:17 | |
| Nisha | I am just putting up the patch for it in some time | 11:17 |
| lucasagomes | Nisha, alright | 11:18 |
| *** Keedya has quit IRC | 11:20 | |
| alineb | could anyone please take a look at https://review.openstack.org/#/c/289256 ? It’s a follow-up patch for the --json option. Thanks! | 11:21 |
| lucasagomes | Nisha, mind if I take a stab on that? I have want to test the agent_ partition images anyway | 11:27 |
| Nisha | yes sure | 11:32 |
| Nisha | lucasagomes, so i need not add a patch fot that ? | 11:33 |
| Nisha | if you are taking care of it? | 11:33 |
| lucasagomes | Nisha, unless you already started it | 11:33 |
| Nisha | I started, but you can do. I have a very minor bug to fix in IPA and ironci for partition images support in agent drivers | 11:34 |
| *** sturivnyi has joined #openstack-ironic | 11:34 | |
| openstackgerrit | Tan Lin proposed openstack/ironic: Add require_exclusive_lock decorators to conductor methods https://review.openstack.org/295734 | 11:35 |
| Nisha | lucasagomes, i propose it it will help if u review it | 11:35 |
| lucasagomes | Nisha, cool sure! | 11:35 |
| lintan | hi jroll, jivillal and zhenguo_, I submit a patch to fix the log request issue for inspector, please take a look at it when you have time: https://review.openstack.org/#/c/295663/ | 11:39 |
| *** dprince has joined #openstack-ironic | 11:45 | |
| *** libu has joined #openstack-ironic | 11:47 | |
| div | still facing swift issue during devstack | 11:50 |
| div | openstack object store account set --property Temp-URL-Key=password 2016-03-21 09:36:49.425 | Not Found (HTTP 404) | 11:50 |
| div | can anybody help | 11:51 |
| *** jcoufal has joined #openstack-ironic | 11:52 | |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Update authorized_keys with new key only https://review.openstack.org/295293 | 11:54 |
| *** libu has quit IRC | 12:01 | |
| alineb | lucasagomes: Thanks for your review! Should I update the commit message? (sorry, i am not sure about how this works after the gate jobs have started...) | 12:03 |
| lucasagomes | alineb, I think it's fine, don't worry | 12:04 |
| alineb | lucasagomes: ok, thank you! | 12:04 |
| lucasagomes | alineb, thank you for the patch (-: | 12:05 |
| *** trown|outtypewww is now known as trown | 12:05 | |
| *** moshele has quit IRC | 12:06 | |
| *** moshele has joined #openstack-ironic | 12:06 | |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic-inspector: Better error handling when converting eDeploy data https://review.openstack.org/295327 | 12:08 |
| *** baoli has joined #openstack-ironic | 12:10 | |
| openstackgerrit | Nisha Agarwal proposed openstack/ironic: Add disk_label for partition images for agent drivers https://review.openstack.org/295751 | 12:18 |
| *** smoriya_ has quit IRC | 12:18 | |
| Nisha | lucasagomes, ^^^ | 12:21 |
| trown | lucasagomes: would you mind looking at https://review.openstack.org/#/c/286070/ last barrier to lighting the bash deploy ramdisk on fire | 12:21 |
| Nisha | IPA patch on the way | 12:21 |
| *** lazy_prince has quit IRC | 12:21 | |
| lucasagomes | Nisha, trown cool, I will take a look in a sec | 12:23 |
| trown | thanks :) | 12:23 |
| Nisha | :) | 12:23 |
| openstackgerrit | Merged openstack/python-ironicclient: Improve output of --json option https://review.openstack.org/289256 | 12:25 |
| *** afaranha has joined #openstack-ironic | 12:26 | |
| *** raildo-afk is now known as raildo | 12:30 | |
| *** mtanino has joined #openstack-ironic | 12:30 | |
| *** lazy_prince has joined #openstack-ironic | 12:33 | |
| *** josh has joined #openstack-ironic | 12:37 | |
| *** josh is now known as Guest99149 | 12:38 | |
| *** killer_prince has joined #openstack-ironic | 12:38 | |
| *** Goneri has joined #openstack-ironic | 12:39 | |
| *** killer_prince has quit IRC | 12:40 | |
| *** vdrok has quit IRC | 12:42 | |
| *** vdrok has joined #openstack-ironic | 12:43 | |
| openstackgerrit | Nisha Agarwal proposed openstack/ironic-python-agent: Add disk_label support for partition images https://review.openstack.org/295766 | 12:46 |
| Nisha | lucasagomes, IPA patch disk_label ^^^ | 12:47 |
| *** moshele has quit IRC | 12:49 | |
| *** moshele has joined #openstack-ironic | 12:50 | |
| openstackgerrit | Gonéri Le Bouder proposed openstack/ironic-python-agent: iscsi: wipe part table before starting the target https://review.openstack.org/284347 | 12:55 |
| *** moshele has quit IRC | 13:00 | |
| *** links has quit IRC | 13:01 | |
| *** dims_ has joined #openstack-ironic | 13:02 | |
| *** dims has quit IRC | 13:02 | |
| *** lazy_prince has quit IRC | 13:07 | |
| *** rbudden has joined #openstack-ironic | 13:07 | |
| *** lucasagomes is now known as lucas-hungry | 13:07 | |
| *** daemontool has quit IRC | 13:07 | |
| *** lazy_prince has joined #openstack-ironic | 13:08 | |
| *** daemontool has joined #openstack-ironic | 13:09 | |
| openstackgerrit | Merged openstack/pyghmi: Have ipv6 addresses always be represented as list https://review.openstack.org/295506 | 13:10 |
| *** baoli has quit IRC | 13:15 | |
| *** chopmann has joined #openstack-ironic | 13:16 | |
| *** baoli has joined #openstack-ironic | 13:16 | |
| *** moshele has joined #openstack-ironic | 13:17 | |
| *** mtanino has quit IRC | 13:17 | |
| jroll | Nisha: so, the agent partition images work is not done?? | 13:26 |
| jroll | morning everyone | 13:27 |
| sambetts | o/ jroll | 13:28 |
| *** Guest99149 has quit IRC | 13:28 | |
| *** baoli has quit IRC | 13:40 | |
| *** baoli has joined #openstack-ironic | 13:41 | |
| *** ametts has joined #openstack-ironic | 13:44 | |
| jroll | lintan: idk if you're around - mind if I update https://review.openstack.org/#/c/238008/10? | 13:45 |
| *** alexpilotti has joined #openstack-ironic | 13:47 | |
| *** daemontool has quit IRC | 13:48 | |
| sambetts | gah! gerrit is really annoying me right now ... I'm trying to leave a comment and everytime I type a word it scrolls down the page for no reason... | 13:54 |
| *** daemontool has joined #openstack-ironic | 13:54 | |
| *** spandhe has joined #openstack-ironic | 13:56 | |
| TheJulia | sambetts: I've found I just have to reload the page when stuff like that starts happeneing :( | 13:56 |
| TheJulia | also found it won't let me scroll up | 13:56 |
| *** ohamada has quit IRC | 13:56 | |
| sambetts | :( | 13:56 |
| sambetts | reload didn't help :( | 13:57 |
| TheJulia | :( | 13:58 |
| *** mgoddard_ has joined #openstack-ironic | 13:59 | |
| *** spandhe has quit IRC | 14:00 | |
| *** ohamada has joined #openstack-ironic | 14:00 | |
| *** mtanino has joined #openstack-ironic | 14:01 | |
| *** mgoddard has quit IRC | 14:03 | |
| zigo | Hi over here! | 14:05 |
| zigo | I have an issue when building the sphinx-doc for Ironic. | 14:05 |
| zigo | http://paste.debian.net/418062/ | 14:05 |
| zigo | (when building the Debian package for 5.0.0) | 14:05 |
| zigo | What's going on? | 14:05 |
| *** links has joined #openstack-ironic | 14:05 | |
| jroll | zigo: interesting... | 14:05 |
| NobodyCam | Good morning Ironicers | 14:06 |
| TheJulia | good morning NobodyCam | 14:06 |
| jroll | zigo: it certainly works in the gate :/ | 14:06 |
| zigo | jroll: Do you have a "it works in devstack" t-shirt? :) | 14:06 |
| zigo | :P | 14:07 |
| TheJulia | lol | 14:07 |
| *** cdearborn has joined #openstack-ironic | 14:07 | |
| jroll | zigo: what I mean is, it looks like a code bug, but it works elsewhere | 14:07 |
| zigo | jroll: Sure, just trying to be funny... :P | 14:07 |
| jroll | there's some magic here https://github.com/openstack/ironic/blob/master/ironic/objects/__init__.py | 14:07 |
| jroll | so maybe that method isn't getting called for whatever reason | 14:08 |
| jroll | zigo: what command are you using to build it? | 14:08 |
| *** Nisha has quit IRC | 14:08 | |
| zigo | jroll: PYTHONPATH=$(CURDIR) sphinx-build doc/source $(CURDIR)/debian/ironic-doc/usr/share/doc/ironic-doc/html | 14:08 |
| NobodyCam | good morning TheJulia jroll jlvillal devananda mgould sinval sambetts lucas-hungry dtantsur|pto gabriel-bezerra and everyone else not directly listed here :) | 14:08 |
| *** achanda has joined #openstack-ironic | 14:09 | |
| jroll | hrm | 14:09 |
| sambetts | o/ NobodyCam | 14:09 |
| zigo | Quite standard ... | 14:09 |
| TheJulia | shouldn't the doc be built in the venv? | 14:09 |
| zigo | TheJulia: That's in the context of building the Debian package, so no venv, no tox, no pip... | 14:10 |
| TheJulia | hmmm | 14:10 |
| *** achanda has quit IRC | 14:10 | |
| mgould | NobodyCam, TheJulia jroll sambetts zigo morning | 14:10 |
| jroll | zigo: okay, I can reproduce that over here | 14:11 |
| zigo | Ah, good! :P | 14:11 |
| jroll | zigo: I'm wondering what the difference is between that command and what 'setup.py build_sphinx' does | 14:11 |
| NobodyCam | :) | 14:11 |
| zigo | Oh, I wonder why I don't have -b html ... | 14:12 |
| *** baoli has quit IRC | 14:12 | |
| sambetts | o/ mgould | 14:13 |
| *** baoli has joined #openstack-ironic | 14:13 | |
| * zigo tries again with -b html | 14:13 | |
| zigo | Same stuff, obviously... | 14:14 |
| jroll | zigo: so nova has similar code, does this same command work for nova? | 14:14 |
| jroll | (code meaning import magic) | 14:14 |
| *** vinm213 has quit IRC | 14:14 | |
| zigo | Yup. | 14:14 |
| zigo | For nova, I do: | 14:14 |
| zigo | sphinx-build -b html doc/source $(CURDIR)/debian/nova-doc/usr/share/doc/nova-doc/html | 14:14 |
| zigo | I do this kind of stuff for maybe 99.99% of all of OpenStack ! :) | 14:15 |
| jroll | hrm | 14:15 |
| jroll | zigo: anyway, could you please file a bug, this is going to take some digging | 14:16 |
| zigo | Ok, doing it now. | 14:19 |
| *** dims_ has quit IRC | 14:19 | |
| jroll | thanks | 14:19 |
| *** lucas-hungry is now known as lucasagomes | 14:21 | |
| lucasagomes | jroll, NobodyCam zigo TheJulia morning | 14:21 |
| TheJulia | good morning lucasagomes | 14:21 |
| zigo | https://bugs.launchpad.net/ironic/+bug/1560508 | 14:21 |
| openstack | Launchpad bug 1560508 in Ironic "Cannot build sphinx doc in Debian for the 5.0.0 release" [Undecided,New] | 14:21 |
| zigo | lucasagomes: Hi ! | 14:21 |
| *** achanda has joined #openstack-ironic | 14:23 | |
| openstackgerrit | Zhenguo Niu proposed openstack/ironic: [Devstack]Add ability to enable shellinabox SSL certificate https://review.openstack.org/289671 | 14:23 |
| *** moshele has quit IRC | 14:24 | |
| *** div has quit IRC | 14:27 | |
| *** mgould has quit IRC | 14:28 | |
| jroll | \o lucasagomes | 14:30 |
| *** openstackgerrit has quit IRC | 14:33 | |
| *** openstackgerrit has joined #openstack-ironic | 14:34 | |
| *** mgould has joined #openstack-ironic | 14:35 | |
| *** jaybeale has joined #openstack-ironic | 14:35 | |
| *** jaybeale has quit IRC | 14:35 | |
| *** ohamada has quit IRC | 14:35 | |
| jlvillal | Good morning NobodyCam jroll zigo lucasagomes sambetts mgould TheJulia and everyone else :) | 14:39 |
| sergek | o/ jlvillal | 14:40 |
| jlvillal | Hi sergek :) | 14:40 |
| TheJulia | good morning jlvillal | 14:41 |
| *** e0ne has quit IRC | 14:41 | |
| jroll | zigo: this wfm https://review.openstack.org/#/c/295868/ | 14:41 |
| *** baoli has quit IRC | 14:42 | |
| zigo | jroll: Thanks a lot, testing it right away. | 14:42 |
| jroll | thanks | 14:42 |
| *** baoli has joined #openstack-ironic | 14:42 | |
| *** links has quit IRC | 14:42 | |
| jroll | quite the silly bug, still am not sure why we don't see it in the gate environment :/ | 14:43 |
| jroll | zigo: I also noticed there's some warnings that go away with -a, you may want to use that? | 14:43 |
| *** achanda has quit IRC | 14:43 | |
| zigo | Ok. | 14:43 |
| *** absubram has joined #openstack-ironic | 14:46 | |
| *** moshele has joined #openstack-ironic | 14:47 | |
| zigo | jroll: This fixes it for me, thanks again, voting +1. | 14:47 |
| *** lazy_prince has quit IRC | 14:48 | |
| jroll | zigo: nice, ty | 14:48 |
| jroll | lucasagomes: mind reviewing 295868? it's pretty nasty, idk if there's a better way though | 14:49 |
| zigo | Uploaded Ironic 5.0.0 to Debian Experimental. | 14:49 |
| jroll | woot, thanks! | 14:49 |
| lucasagomes | jroll, will do in a sec (I'm in a call) | 14:49 |
| zigo | FYI, it will go to Unstable (and therefore testing) with the rest of Mitaka when then we have a final release. | 14:49 |
| *** jaybeale has joined #openstack-ironic | 14:49 | |
| *** ohamada has joined #openstack-ironic | 14:49 | |
| jroll | lucasagomes: no worries | 14:50 |
| jroll | zigo: we've got a 5.1.0 coming this week for the final mitaka release, jfyi | 14:51 |
| zigo | Good to know. | 14:51 |
| *** ayoung has joined #openstack-ironic | 14:51 | |
| zigo | Will there be a new ironic-inspector too? | 14:51 |
| zigo | Or just 3.2.0 like I just uploaded? | 14:52 |
| TheJulia | I believe, based on the notes I saw yesterday that 3.2.0 is intended to be the mitaka release for inspector | 14:52 |
| zigo | ok | 14:52 |
| ayoung | Hey guys...long time listener, first time caller. Love the show. Question about IPMI based compute node bringup. Is it possible to inject a file ( a secret key) per node when doing openstack server create ? | 14:53 |
| jroll | zigo: TheJulia: correct, 3.2.0 is mitaka ironic-inspector | 14:53 |
| jroll | zigo: we just released ironic-python-agent 1.2.0 as mitaka, ironic, ironic-ui, and bifrost are coming this week | 14:53 |
| sambetts | ayoung: http://docs.openstack.org/user-guide/cli_provide_user_data_to_instances.html | 14:53 |
| TheJulia | ayoung: Awesome! So, openstack server create as in the OSC command line? | 14:54 |
| zigo | ironic-python-agent? | 14:54 |
| ayoung | sambetts, ok, so how does that data get down to the node? Is it secure? | 14:54 |
| zigo | I didn't know about it. | 14:54 |
| jroll | zigo: yeah, dunno if you package that, IMO it doesn't need to be packaged | 14:54 |
| zigo | I didn't know there was a plugin for horizon either. | 14:54 |
| ayoung | TheJulia, yeah...use it in Tripleo. I can get you the exact command... | 14:54 |
| zigo | Do you know if ironic-ui is Django 1.9 ready? | 14:54 |
| TheJulia | zigo: the horizon plugin is in early development | 14:54 |
| zigo | I can't upload it if it's not. | 14:54 |
| ayoung | TheJulia, http://adam.younglogic.com/2016/03/host-tripleo-overcloud/ | 14:54 |
| sambetts | ayoung: the file is either loaded using the nova metadata service or config drive | 14:55 |
| *** mgoddard_ has quit IRC | 14:55 | |
| TheJulia | zigo: betherly ^^^ | 14:55 |
| zigo | Horizon gained Django 1.9 compat since yesterday, so I could finally upload it to Debian. | 14:55 |
| *** mgoddard has joined #openstack-ironic | 14:55 | |
| jroll | zigo: ironic-python-agent is a python app to destroy (or image!) a machine, intended to run in a ramdisk. IMO I never want to apt-get install that :) | 14:55 |
| zigo | I already started fixing some of the plugins (I have 2 patches for sahara-dashboard for example). | 14:55 |
| jroll | zigo: all ironic-ui questions may be directed to betherly | 14:55 |
| zigo | Ok. | 14:55 |
| zigo | jroll: FYI, for running in a ramdisk, I use debian-live ! :) | 14:56 |
| zigo | It works super well. | 14:56 |
| ayoung | sambetts, so metadata server is out. Its essentially a public website, visible by everything on the metadata network, so all Ironic hosts. | 14:56 |
| zigo | That's how I run tempest functional tests for my packages. | 14:56 |
| jroll | sambetts: ayoung: if using configdrive, it is not terribly secure - the configdrive is stored in either ironic's database or swift, not encrypted | 14:56 |
| zigo | Re-image? Just reset ... :P | 14:56 |
| TheJulia | ayoung: what sambetts said, that being said, if a configuration drive is used, it can only be trusted on the first boot-up of a machine | 14:56 |
| jroll | zigo: neat, we use coreos with a debian chroot | 14:56 |
| TheJulia | and could be modified by a sufficently permissioned user after the fact once on disk... | 14:57 |
| TheJulia | tl;dr, putting secret files in config drives is generally not a great idea | 14:57 |
| jroll | metadata service can be secure, fwiw, if your network is secure (as I understand it) | 14:58 |
| ayoung | jroll, so, we only need to put a one-time-password in it. If it is stored unencrypted it is ok, so long as not "everyone" can fetch it. | 14:58 |
| jroll | ayoung: right, so only ironic operators and people with access to that server can fetch it | 14:58 |
| ayoung | I have to admit I don't 100% understand the Ironic boot process. I assume it starts with an IPMI wakeup...I thought it would be PXE, but I guess not? | 14:58 |
| *** saripurigopi has joined #openstack-ironic | 14:59 | |
| TheJulia | ayoung: depends on how the node is deployed, if it is a whole disk image, partition image, if the node is requested to always netboot | 14:59 |
| jroll | ayoung: let me find you a doc :) | 14:59 |
| *** lazy_prince has joined #openstack-ironic | 14:59 | |
| *** jaosorior has joined #openstack-ironic | 14:59 | |
| *** izaakk has joined #openstack-ironic | 15:00 | |
| ayoung | jroll, thanks. This "initial safe enroll" thing has been a moving target. | 15:00 |
| jroll | ayoung: these two diagrams may be helpful http://docs.openstack.org/developer/ironic/deploy/user-guide.html#example-1-pxe-boot-and-iscsi-deploy-process | 15:00 |
| *** rcrit_ has joined #openstack-ironic | 15:00 | |
| ayoung | rcrit_, last line was | 15:01 |
| ayoung | ayoung: these two diagrams may be helpful http://docs.openstack.org/developer/ironic/deploy/user-guide.html#example-1-pxe-boot-and-iscsi-deploy-process | 15:01 |
| jroll | ayoung: there are some other drivers that use virtualmedia instead of pxe | 15:01 |
| jroll | and attach an iso through that channel | 15:01 |
| ayoung | so PXE does a tftp fetch, which does not really have any security in it. I was hoping IPMI would drop a key there that could be used as part of the kernel boot or something... | 15:01 |
| ayoung | But I guess it is just a "dumb" power on. | 15:02 |
| ayoung | conductor could, in theory, inject a keypair into the image it returns in the tftp response | 15:03 |
| jroll | right, so | 15:03 |
| jroll | ipmi is dumb, it can't do anything fun | 15:03 |
| jroll | some hardware with virtualmedia can securely attach the image | 15:03 |
| jroll | however | 15:03 |
| jroll | configdrive is always transferred via iscsi or http | 15:04 |
| jroll | tftp is only the deploy ramdisk (except in vmedia case) and the kernel/ramdisk of the user image in most cases | 15:04 |
| *** alexpilotti has quit IRC | 15:07 | |
| rcrit_ | jroll, what about the nova injected files? | 15:07 |
| *** keedya has joined #openstack-ironic | 15:07 | |
| ayoung | jroll, there were some security extension for IPMI, but I was not clear if they were Cisco specific\ | 15:08 |
| *** suro-patz has joined #openstack-ironic | 15:08 | |
| ayoung | rcrit_, config drive | 15:08 |
| jroll | rcrit_: those go via configdrive | 15:08 |
| ayoung | that was earlier in our conversation. | 15:08 |
| rcrit_ | ok, sorry missed that | 15:08 |
| jroll | ayoung: right, but IPMI cannot inject files, whether it's secure or not | 15:08 |
| ayoung | so that is fetched from the conductor Database (or swift) | 15:08 |
| jroll | and IPMI is insecure as heck | 15:08 |
| rcrit_ | is the nova vendoradata plugin supported? | 15:08 |
| rcrit_ | to provide custom cloud-init scripts? | 15:08 |
| jroll | I believe so yeah, we inject vendor data in our deployment | 15:09 |
| jroll | nova builds the configdrive as usual, we just write it to a partition | 15:10 |
| ayoung | so that is still configdrive | 15:11 |
| sambetts | If your running the metadata service user data or vendor data can be downloaded from there also | 15:11 |
| rcrit_ | nice | 15:11 |
| ayoung | what I would really like is a way that a machine could do a post "here is my public key" and a way to validate, networkwise, that it came from the right machine | 15:12 |
| ayoung | like, something at the layer 2 level: it came from macaddress fooo and we know that came over the right port. | 15:13 |
| sambetts | well with config drive you don't need that because we directly write the data to the machines disk | 15:14 |
| ayoung | sambetts, oh? | 15:14 |
| lucasagomes | jroll, uu :-( odd that it only failing in debian tho | 15:14 |
| jaosorior | ayoung: not sure if we could make that work... you could easily do mac spoofing | 15:14 |
| ayoung | sambetts, can we lock it down so that only that machine gets it somehow? | 15:14 |
| ayoung | jaosorior, yeah, yeah...that is why the port check...but I am not inventing a security algorthm here. I know better | 15:14 |
| ayoung | just dreaming | 15:14 |
| jaosorior | ayoung, rcrit_: We fall back to the issue we were dealing with before. the contents of the injected files and the network info are gonna be available in the metadata service | 15:15 |
| jroll | lucasagomes: yeah, weird | 15:15 |
| sambetts | ayoung: thats what happens anyway, during the deploy process, at the same time as we write the glance image to the node, we also write a config drive partition specificly generated for that node | 15:15 |
| ayoung | sambetts, when you say "we write the data to the machines disk" what protocol is that done over? | 15:15 |
| sambetts | iscsi | 15:16 |
| jroll | well | 15:16 |
| rcrit_ | jaosorior, I think this might be a special case as it is written to disk and available only to a booted machine AFAIU | 15:16 |
| jroll | like I said before, configdrive may be transferred over iscsi, http(s), virtualmedia | 15:16 |
| jroll | depending on hardware/drivers | 15:16 |
| sambetts | jroll: += | 15:16 |
| sambetts | ++ | 15:16 |
| *** achanda has joined #openstack-ironic | 15:17 | |
| *** garthb has joined #openstack-ironic | 15:17 | |
| ayoung | right, so we could only "guarantee" the delivery in an iscsi TLS case | 15:17 |
| jaosorior | rcrit_ https://github.com/openstack/nova/blob/master/nova/virt/ironic/driver.py#L653 | 15:17 |
| jroll | ? | 15:17 |
| jroll | https isn't secure enough? | 15:18 |
| jroll | or virtualmedia over tls? | 15:18 |
| ayoung | jroll, actually, for this usage, https should be sufficient, too. | 15:18 |
| jroll | right | 15:18 |
| jroll | ayoung: I don't believe we support iscsi over TLS today, but I may be wrong | 15:18 |
| ayoung | I more worried about passive evesdropping than a man in the middle for the boot process. | 15:19 |
| ayoung | I'm | 15:19 |
| jroll | sure | 15:19 |
| *** sinval has quit IRC | 15:20 | |
| jroll | actually I don't think the agent can listen for tls yet either, that throws that out | 15:20 |
| *** harlowja_at_home has joined #openstack-ironic | 15:20 | |
| jroll | honestly if you're going for max security, I'd go with ilo + virtualmedia + swift driver | 15:20 |
| jaosorior | jroll: Why not? | 15:20 |
| * jroll findds link | 15:20 | |
| jroll | jaosorior: it would have to be a client cert situation, you'd have thousands of these agents and they're super ephemeral | 15:21 |
| jaosorior | I see | 15:21 |
| sambetts | ayoung: is there a reason your running your Ironic provisioning network on your public network? can you not make the machine accessible via a natted floating ip? | 15:21 |
| ayoung | sambetts, assume a node is compromised, and all bets are off | 15:21 |
| jaosorior | well, the provisioning network shouldn' | 15:22 |
| jaosorior | shouldn't be in the public network | 15:22 |
| ayoung | so, this is the provisioning network. | 15:22 |
| jaosorior | it should be separate from the external and even the internal API networks | 15:22 |
| * jaosorior talking about TripleO | 15:22 | |
| * ayoung trying to replace the screendoor on his submarine. | 15:23 | |
| ayoung | sambetts, so, we are trying to come up with an enrollment process for all the deployments, and just trying to figure out the rules of the game. | 15:24 |
| TheJulia | what do you mean by enrollment? | 15:24 |
| ayoung | We were just looking at Nova and booting image there, but realized that Ironic, due to its variety of boot processes, might have different rules | 15:24 |
| ayoung | TheJulia, I mean "how do you identify a server?" | 15:25 |
| TheJulia | ayoung: identify an already known server? | 15:25 |
| ayoung | TheJulia, so, say you want to use ssh to a machine, how do you know you have the right machine? | 15:25 |
| *** suro-patz has quit IRC | 15:25 | |
| TheJulia | true, so we should avoid the word enrollment in that case | 15:25 |
| TheJulia | since we have an enrollment state that new nodes are added into | 15:26 |
| jroll | let's step back a moment | 15:26 |
| ayoung | TheJulia, well, we mean it as short for "enroll an client into the Identity management system" | 15:26 |
| jroll | ayoung: how would you do this for a vm? | 15:26 |
| TheJulia | ayoung: ahh | 15:26 |
| jaosorior | jroll, sambetts: dumb question; Once the injected files are persisted to the configdrive. Can one still access that data via the metadata service? Or does it not get persisted there? | 15:26 |
| ayoung | jroll, we actually have a lot of the same problems there | 15:26 |
| ayoung | jroll, rcrit_ had a proof of concept working with nova hooks, but really, there was no good solution yet | 15:27 |
| jroll | ayoung: yeah, so from the nova user perspective we intend to behave like any nova VM (with some extra horsepower and a few caveats) | 15:27 |
| ayoung | I was looking into securing the message queue and getting the information that way, but with Ironic, there is no Queue listener on the node. | 15:27 |
| ayoung | jroll, and that assumes that the compute node itself can be trusted | 15:27 |
| ayoung | and...well, how do we provision compute nodes...ironic... | 15:28 |
| ayoung | and thus we are bothering you and flooding your chat room this morning | 15:28 |
| jroll | jaosorior: yes, configdrive == metadata service, in terms of data. delivery method is the only variance. | 15:28 |
| jroll | ayoung: right, so solving this from the nova user perspective (for both virt and bare metal) is the right thing to do, in my mind | 15:28 |
| rcrit_ | but in this case the metadata service is local to the box right? We're not talking metadata service over HTTP right? | 15:29 |
| ayoung | rcrit_, nope. It is on the provisioning server | 15:29 |
| jroll | ayoung: because your undercloud is just another nova deployment (that happens to use the ironic driver) | 15:29 |
| ayoung | jroll, If I can secure the message queue, I can solve it for the overcloud | 15:29 |
| ayoung | for the undercloud, I need a way to secure the initial bringup process | 15:30 |
| ayoung | And, it sounds like a fools errand at this point | 15:30 |
| jroll | ayoung: only if you are an operator with access to the queue no? | 15:30 |
| ayoung | I'm used to those. I work on Keystone | 15:30 |
| jroll | ayoung: anyway, if you can secure the undercloud's message queue, you can use the same method, right? | 15:30 |
| ayoung | jroll, not quite. In the undercloud, there is no message listener on the host being brought up. Its all PXE type stuff | 15:31 |
| jroll | ayoung: well, we still have nova-compute hosts | 15:31 |
| ayoung | overcloud is only as secure as the undercloud. Undercloud is only as secure as the hardware provisiong | 15:31 |
| jroll | I guess I'd need to understand the solution involving the queue | 15:31 |
| jroll | before I can say if that would work | 15:32 |
| *** vishwanathj_zzz is now known as vishwanathj | 15:32 | |
| ayoung | jroll, in the case of the overcloud, the openstack compute process is running on the hypervisor node | 15:32 |
| ayoung | mayeb not in a vsphere deployment, but they have their own security from there on down anyway | 15:32 |
| ayoung | but for Centos/libvirt | 15:32 |
| jroll | ok sure, and the end goal is instance ssh host keys being passed back to the nova user, right? | 15:33 |
| ayoung | so you have a secure way of delivering a message to the compute. I am not saying, by the way, that this is a solved problem | 15:33 |
| ayoung | jroll, right | 15:33 |
| jroll | so how does the hyp get the host keys? | 15:33 |
| ayoung | jroll, here is what I would love to see, if I were dreaming | 15:33 |
| ayoung | on the outside of every machine I rack, it had a public key printed that I could scan in using a barcode reader | 15:34 |
| ayoung | when it pxebooted, part of the request includes a Certificate Signing Request or something comparable | 15:34 |
| ayoung | that CSR would be compared with the public key off the server, and, bang, I know I have the right server. | 15:35 |
| ayoung | jroll, private key never leaves the newly provisioned host | 15:35 |
| jroll | right, so I'm even more confused now | 15:35 |
| ayoung | jroll, its ok, it took me years to get this warped | 15:36 |
| ayoung | jroll, the thing that I am looking for is a way to trust a machine from as early in the install process as possible | 15:36 |
| devananda | morning, all | 15:36 |
| jroll | but AIUI, all you really need is for the ssh (public) host key that are within the already provisioned bare metal node, to be passed back to nova for the api user to fetch | 15:36 |
| ayoung | and to avoid having secrets go across the network if possible. Just trying to get at the assumptions | 15:36 |
| ayoung | jroll, right...when I say CSR, that is really the same thing. | 15:37 |
| ayoung | the ssh public key | 15:37 |
| jroll | right. | 15:37 |
| jroll | so I believe that this is a nova problem, that should be solved for all nova instances | 15:37 |
| sambetts | during the install process the node is completely wiped and overwritten so how could you store anything on it unless something hardware specific ? | 15:38 |
| jroll | and honestly probably does not depend on the virt driver, as no virt driver should have access to read keys from the instance itself | 15:38 |
| ayoung | jroll, so, yeah. If cloud-init kicked off an ssh-keygen, then the question would be "how do we trust that key K is from server S?" | 15:38 |
| jroll | the only way I really see to do that is auth by IP address | 15:39 |
| TheJulia | create an api where the public key can be posted someplace along with a profile of the hardware | 15:39 |
| jroll | and make damn sure your network is secure enough that IPs cannot be spoofed | 15:39 |
| ayoung | TheJulia, I think that is "necessary" | 15:39 |
| ayoung | jroll, hence me muttering about layer 2 and all that earlier. Now I am warping you... | 15:39 |
| *** awiddersheim has joined #openstack-ironic | 15:39 | |
| TheJulia | it is similar to what occurs when IPA phones home, we compare the hardware and then record the node's IP address to call back to | 15:39 |
| jroll | ayoung: there's no way we can run code inside a nova instance, whether virt or bare metal | 15:39 |
| openstackgerrit | Gonéri Le Bouder proposed openstack/ironic-python-agent: iscsi: wipe part table before starting the target https://review.openstack.org/284347 | 15:39 |
| ayoung | jroll, we can kcik something off with cloud-init | 15:40 |
| sambetts | jroll: unless they want to inject a script via config-drive or metadata | 15:40 |
| TheJulia | ayoung: this really sounds like a cloud-init kind of feature | 15:40 |
| jroll | ayoung: unless it is injected into the image/configdrive, which is not an ironic-specific problem | 15:40 |
| ayoung | jroll, right...which is why we were looking in to how is the configdrive delivered | 15:40 |
| ayoung | right now that sounds like the best option | 15:40 |
| * devananda catches up on the conversation | 15:40 | |
| jroll | ayoung: well, I still think it's bad to pass secrets in the configdrive (whether vm or bm), because the operator can access that | 15:41 |
| jroll | (in the general case) | 15:41 |
| TheJulia | ayoung: but that means your centralizing that potentially, just seems a little more risky if that host is compromised | 15:41 |
| jroll | ++ | 15:41 |
| TheJulia | the phone home concept, seems the most sane | 15:42 |
| *** alexpilotti has joined #openstack-ironic | 15:42 | |
| rcrit_ | but phone home with what? How do you prove possession? | 15:42 |
| sambetts | mac address? | 15:42 |
| ayoung | jroll, so, when I did my first proof of concept, the flow was like this: 1. Contact the identity provider and create a host entry, including a OTP (one time password). 2. Call nova boot with the OTP in the user-data. 3. Have cloud-init call a script that used the OTP to autoregister the instace with the identity provider | 15:42 |
| TheJulia | hardware profile | 15:42 |
| jroll | neither, IP address | 15:43 |
| ayoung | so, while it was on the config drive, the window was small: just during the boot process | 15:43 |
| jroll | and make sure things can't spoof ip addresses | 15:43 |
| *** alexpilo_ has joined #openstack-ironic | 15:43 | |
| rcrit_ | ayoung, the thing is though if we make a general system then someone may put longer-term secrets in there which would be bad | 15:43 |
| TheJulia | eh, I think spoofing the ip would be really easy if one has compromised the environment | 15:43 |
| jroll | ayoung: I guess that works in the undercloud case where you trust your provider, I would not trust that as a public cloud user | 15:44 |
| jroll | TheJulia: not if your ToRs prevent it :) | 15:44 |
| ayoung | rcrit_, agreed. I was the one that origianlly objected. As I said "proof of concept" | 15:44 |
| jlvillal | devananda: I think you did a +2 on this before: https://review.openstack.org/#/c/287306/ Not sure if you could give it a quick look again? | 15:44 |
| jroll | if you've compromised the network it's all over anyway | 15:44 |
| devananda | ayoung: as jroll said previously, there are mechanisms in place to securely pass data into the Ironic instance, if the hardware and the driver supports virtual media | 15:44 |
| devananda | ayoung: but it also sounds like you're trying to solve a special case (cloud operator deploying cloud resources) within the constraints of a general case (cloud user wishes to trust the cloud but can't control the boot process) | 15:44 |
| TheJulia | jroll: what if I did it in the controlplane? *ducks* | 15:44 |
| ayoung | jroll, so, if we can secure the undercloud, we could then deliver secrets via the message queue (assume we secure that) | 15:44 |
| ayoung | when deplouying in the overcloud | 15:44 |
| NobodyCam | morning devananda | 15:45 |
| ayoung | yeah, soundsl ike the virtualmedia approach is the best. Looks like that is tied to ilo? | 15:45 |
| jroll | ayoung: right, I think your best option today, for tripleo's case, is what you're doing, and use virtualmedia | 15:45 |
| jroll | I believe irmc also has it, maybe drac? | 15:45 |
| *** dims has joined #openstack-ironic | 15:45 | |
| devananda | ayoung: I fail to see why you're constraining the securitization of the undercloud deployment methods within the untrusted framework of a public cloud or shared network environment | 15:45 |
| openstackgerrit | Sivaramakrishna Garimella proposed openstack/ironic: Update the deploy drivers with network flipping logic https://review.openstack.org/213262 | 15:45 |
| *** sivaramakrishna has joined #openstack-ironic | 15:46 | |
| ayoung | devananda, let me try and parse that... | 15:46 |
| jaosorior | devananda: Cause this solution is not only meant for the undercloud deployment | 15:46 |
| devananda | jaosorior: ahh, ok | 15:46 |
| jroll | well, I don't think the solution of posting passwords to the identity service is a good solution for the public cloud case | 15:46 |
| jroll | s/passwords/secrets/ | 15:47 |
| *** alexpilotti has quit IRC | 15:47 | |
| ayoung | devananda, ok, so If i understand the problem correctly, the security of any layer is dependent on the layer below it. We can't securely identify things in the overcloud if we can't already do that in the undercloud | 15:47 |
| jroll | then again, maybe it's "good enough" | 15:47 |
| sambetts | jroll: I think they would be posting their pub key and therefore never transfering the private key on the network | 15:47 |
| ayoung | so I am trying to push it back as early in the process as we can, to the provisioning of the bare metal machines | 15:47 |
| *** Nisha_away has joined #openstack-ironic | 15:47 | |
| jaosorior | posting secrets to the identity service? jroll, are you talking about the OTPs for FreeIPA that ayoung mentioned? | 15:47 |
| jroll | jaosorior: yep | 15:48 |
| jroll | like I said, probably close enough | 15:48 |
| devananda | ayoung: yes. however, the constraints placed upon a public cloud are different -- a user is not also a privileged operator | 15:48 |
| *** alexpilo_ has quit IRC | 15:48 | |
| devananda | ayoung: whereas in the undercloud, AIUI, that same level of privilege separation does not apply | 15:48 |
| *** Sukhdev has joined #openstack-ironic | 15:48 | |
| ayoung | devananda, understood. And maybe the constraints I am placing on the process are too draconian. | 15:49 |
| jroll | I mean, the dream is a perfectly secure deployment | 15:50 |
| ayoung | My assumption, though, is that even the underclouds will grow large enough that securing them will need to be automated. Its one thing when you have 10 machines in one rack, and something different when multisite 100K node deployments are in place | 15:50 |
| jroll | whether that's an ironic public cloud or some undercloud | 15:50 |
| jroll | so "too draconian" isn't ever a thing | 15:50 |
| ayoung | heh | 15:50 |
| *** saripurigopi has quit IRC | 15:50 | |
| devananda | ayoung: securing the deployment process is a thing that has to happen, and yea, I agree with jroll on "there's always going to be room to make it more secure" | 15:51 |
| ayoung | and, if we have soemthing that can work in "some" configuration, but not all, at least we can architect toward that configuration . "Yes, we can do IPMI, but we can only do it secure with ILO or DRAC" for example | 15:51 |
| devananda | ayoung: I believe that ^ is what we've been saying | 15:51 |
| jroll | deployments with IPMI can still be very secure | 15:51 |
| ayoung | so if we say "use config drive to get the OTP on to the newly provisioned machines, and it will be secure if you do x , y, z" it is a start | 15:51 |
| * jroll should do a blog post on "super secure ironic deployments with pizza box servers" | 15:52 | |
| devananda | jroll: ++ | 15:52 |
| TheJulia | jroll: ++ | 15:52 |
| ayoung | jroll, I would love to read that | 15:52 |
| * TheJulia would like the pizza though | 15:52 | |
| jroll | ayoung: the biggest things are: use swift for configdrive transfer, use swift as the backend for images with the agent driver (https), and secure the hell out of your network | 15:53 |
| jroll | use https for ironic api | 15:53 |
| jroll | this should give you encrypted everything except ipmi power control, and lock down that network | 15:53 |
| jroll | there's a few small holes to plug but this gets you most of the way there | 15:54 |
| ayoung | jroll, why swift? | 15:54 |
| ayoung | the rest I get | 15:54 |
| TheJulia | tempurl capability | 15:55 |
| jroll | ayoung: because in the normal case, the configdrive is transferred from ironic -> agent via http, and there isn't a tls mechanism in place there | 15:55 |
| *** dims_ has joined #openstack-ironic | 15:55 | |
| Nisha_away | jroll, following minor patches were raised for partition images for agent drivers https://review.openstack.org/295751, https://review.openstack.org/295766, https://review.openstack.org/295710 | 15:55 |
| *** dims has quit IRC | 15:55 | |
| Nisha_away | these are minor patches | 15:55 |
| jroll | Nisha_away: oh there you are | 15:55 |
| ayoung | jroll, but swift has https? | 15:55 |
| jroll | Nisha_away: I created the stable/mitaka branch last night for IPA | 15:55 |
| *** sivaramakrishna has quit IRC | 15:55 | |
| Nisha_away | oh | 15:55 |
| jroll | ayoung: yes, the problem is the api server on the agent can't do tls today, because certs are hard | 15:56 |
| Nisha_away | jroll, then we dont have disk_label support fo rpartition images for agent drivers in Mitaka :( | 15:56 |
| jroll | Nisha_away: that may be backportable but it'll need to wait | 15:56 |
| devananda | ayoung: yes. and tempurls act similarly to OTP plus data over HTTPS | 15:56 |
| jroll | I understand | 15:56 |
| Nisha_away | ok | 15:56 |
| jroll | Nisha_away: actually, could you file a separate bug for something like "cannot use gpt with partition images in agent driver" for those patches? | 15:57 |
| openstackgerrit | Kyrylo Romanenko proposed openstack/python-ironicclient: Add CLI tests to check JSON response body https://review.openstack.org/291239 | 15:57 |
| Nisha_away | jroll, actually ironic_lib will automatically assign disk_label if nothing is sent while calling work_on_disk() | 15:58 |
| openstackgerrit | Zhenguo Niu proposed openstack/ironic: Restart consoles on conductor startup https://review.openstack.org/257252 | 15:58 |
| Nisha_away | which assigns msdos for bios and gpt for uefi as default disk_labels and those works | 15:58 |
| jroll | Nisha_away: well, what's the issue then? can't specify disk label? | 15:58 |
| jroll | Nisha_away: in any case, please file a bug for the specific bug that this fixes, that's the first step to making this backportable (if it is) | 15:59 |
| Nisha_away | yes, its not sent as an argument to ironci_lib or rather you can say that it doesnt pass user specified disk_label in the node | 15:59 |
| Nisha_away | ok | 15:59 |
| jroll | Nisha_away: thanks | 15:59 |
| Nisha_away | i will file a bug. So in ironic also it will also fixed as a bug then? | 16:00 |
| jroll | I need to step away for a while, back in an huor | 16:00 |
| jroll | Nisha_away: yes, file a single bug against ironic and IPA | 16:00 |
| Nisha_away | jroll, ok, but there is one doc patch for ilo drivers | 16:00 |
| Nisha_away | in among three | 16:01 |
| *** ohamada has quit IRC | 16:01 | |
| jroll | Nisha_away: right, that one should be against the RFE still | 16:01 |
| Nisha_away | Yes | 16:01 |
| Nisha_away | for other two patches i will mark them against the new bug | 16:02 |
| jroll | Nisha_away: okay, ty | 16:02 |
| jroll | bbl | 16:02 |
| openstackgerrit | Sergii Turivnyi proposed openstack/python-ironicclient: Add sanity tests for testing actions with Port https://review.openstack.org/240944 | 16:03 |
| openstackgerrit | Sergii Turivnyi proposed openstack/python-ironicclient: Negative tests for testing actions with port. https://review.openstack.org/259354 | 16:03 |
| *** alexpilotti has joined #openstack-ironic | 16:04 | |
| *** keedya has quit IRC | 16:05 | |
| *** aginwala has joined #openstack-ironic | 16:07 | |
| *** phuongnh has joined #openstack-ironic | 16:07 | |
| *** alexpilotti has quit IRC | 16:09 | |
| *** aginwala has quit IRC | 16:13 | |
| openstackgerrit | Nisha Agarwal proposed openstack/ironic: Add disk_label for partition images for agent drivers https://review.openstack.org/295751 | 16:13 |
| openstackgerrit | Nisha Agarwal proposed openstack/ironic-python-agent: Add disk_label support for partition images https://review.openstack.org/295766 | 16:15 |
| *** jaosorior has quit IRC | 16:18 | |
| *** aginwala has joined #openstack-ironic | 16:18 | |
| *** saripurigopi has joined #openstack-ironic | 16:19 | |
| *** e0ne has joined #openstack-ironic | 16:21 | |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Update resources only for specific instance. https://review.openstack.org/287498 | 16:22 |
| openstackgerrit | Sergii Turivnyi proposed openstack/python-ironicclient: Tests for testing chassis-create command https://review.openstack.org/293634 | 16:26 |
| *** cdearborn has quit IRC | 16:28 | |
| *** aginwala has quit IRC | 16:29 | |
| *** alexpilotti has joined #openstack-ironic | 16:31 | |
| *** moshele has quit IRC | 16:31 | |
| *** aginwala has joined #openstack-ironic | 16:31 | |
| *** chopmann has quit IRC | 16:32 | |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic-python-agent: iscsi: wipe part table before starting the target https://review.openstack.org/284347 | 16:35 |
| *** alexpilotti has quit IRC | 16:36 | |
| *** saripurigopi has quit IRC | 16:37 | |
| jroll | lucasagomes: NobodyCam: jlvillal: TheJulia: I added some things to https://etherpad.openstack.org/p/ironic-mitaka-finish if you want to help dogpile on reviews | 16:41 |
| jlvillal | jroll: Thanks. I will look... | 16:42 |
| lucasagomes | jroll, just +2'd the chassis one | 16:42 |
| lucasagomes | lemme take a look at the others | 16:42 |
| jroll | lucasagomes: thanks! I'll make that a fixme | 16:43 |
| lucasagomes | jroll, not a big deal tho | 16:43 |
| jroll | too late :P | 16:43 |
| lucasagomes | lol | 16:43 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Fix sphinx docs build https://review.openstack.org/295868 | 16:43 |
| * lucasagomes re +2 | 16:44 | |
| jlvillal | jroll is the farmer and I'm a duck :) https://www.youtube.com/watch?v=kjdUSPe7Yj4 | 16:44 |
| openstackgerrit | Ramamani Yeleswarapu proposed openstack/ironic: Centralize config options - [amt] https://review.openstack.org/295946 | 16:45 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Append 'Openstack-Baremetal-Request-ID' header to the response https://review.openstack.org/238008 | 16:45 |
| jroll | jlvillal: O_o | 16:45 |
| jroll | lucasagomes: ^ fixed the response header there | 16:46 |
| *** spandhe has joined #openstack-ironic | 16:46 | |
| * lucasagomes looks | 16:46 | |
| lucasagomes | just finishing the disk label one | 16:46 |
| jlvillal | heh, I always find that video funny | 16:46 |
| *** dims_ has quit IRC | 16:46 | |
| *** suro-patz has joined #openstack-ironic | 16:47 | |
| jroll | jlvillal: in a meeting so I didn't click yet :) | 16:47 |
| *** pcaruana has quit IRC | 16:52 | |
| lucasagomes | inspector folks, can I get some eyes at https://review.openstack.org/#/c/295327/ (small patch) | 16:57 |
| lucasagomes | thanks | 16:57 |
| * sambetts looking | 16:57 | |
| *** keedya has joined #openstack-ironic | 16:57 | |
| *** mgould has quit IRC | 16:58 | |
| sambetts | lucasagomes: +2 | 16:59 |
| lucasagomes | sambetts, ty! | 16:59 |
| *** trown is now known as trown|lunch | 17:01 | |
| *** lucasagomes is now known as lucas-afk | 17:02 | |
| *** Guest91251 has quit IRC | 17:02 | |
| *** dims has joined #openstack-ironic | 17:02 | |
| openstackgerrit | Nisha Agarwal proposed openstack/ironic-python-agent: Add disk_label support for partition images https://review.openstack.org/295766 | 17:04 |
| NobodyCam | night lucas-afk | 17:05 |
| openstackgerrit | Nisha Agarwal proposed openstack/ironic: Add disk_label and node_uuid for agent drivers https://review.openstack.org/295751 | 17:07 |
| phuongnh | sambetts:do you have any idea for me? | 17:08 |
| *** phuongnh has quit IRC | 17:08 | |
| *** baoli has quit IRC | 17:10 | |
| *** baoli has joined #openstack-ironic | 17:11 | |
| openstackgerrit | Gonéri Le Bouder proposed openstack/ironic-python-agent: iscsi: wipe part table before starting the target https://review.openstack.org/284347 | 17:11 |
| *** phuongnh has joined #openstack-ironic | 17:11 | |
| *** harshs has joined #openstack-ironic | 17:12 | |
| jlvillal | lucas-afk: jroll: I did a +2 on https://review.openstack.org/#/c/284347/ I wasn't sure if wanted to only do +A on patches on https://etherpad.openstack.org/p/ironic-mitaka-finish | 17:12 |
| *** dims has quit IRC | 17:13 | |
| jroll | jlvillal: IPA already has the stable/mitaka branch cut, newton is open for it | 17:13 |
| *** mgould has joined #openstack-ironic | 17:13 | |
| jroll | the other projects, don't want to land anything risky | 17:13 |
| jlvillal | jroll: Thanks! | 17:14 |
| jlvillal | jroll: Never mind Goneri pushed a new patch over top of two +2s ;) | 17:15 |
| jroll | heh | 17:15 |
| *** degorenko is now known as _degorenko|afk | 17:15 | |
| *** saripurigopi has joined #openstack-ironic | 17:16 | |
| *** dprince has quit IRC | 17:17 | |
| *** aginwala has quit IRC | 17:17 | |
| *** baoli has quit IRC | 17:18 | |
| *** aginwala has joined #openstack-ironic | 17:18 | |
| *** MattMan has quit IRC | 17:19 | |
| *** MattMan has joined #openstack-ironic | 17:19 | |
| *** aginwala has quit IRC | 17:20 | |
| *** baoli has joined #openstack-ironic | 17:21 | |
| sambetts | o/ jlvillal | 17:22 |
| *** daemontool has quit IRC | 17:22 | |
| jlvillal | sambetts: \o :) | 17:23 |
| *** cdearborn has joined #openstack-ironic | 17:23 | |
| Goneri | jlvillal, yes, there was this ugly type in the docstring. | 17:24 |
| jlvillal | Goneri: I thought lucas-afk had pushed a fix for it? | 17:24 |
| jlvillal | In patch set 19 | 17:25 |
| openstackgerrit | Merged openstack/ironic: Update the text in user guide of ironic https://review.openstack.org/279503 | 17:25 |
| Goneri | oh indeed | 17:25 |
| Goneri | Well, let's pretend that two push in a row give better result :D | 17:26 |
| jlvillal | Goneri: Any thoughts on my minor comment in patch set 19? | 17:27 |
| *** mbound has quit IRC | 17:27 | |
| Goneri | oh indeed, let's fix that. | 17:27 |
| *** krtaylor has quit IRC | 17:27 | |
| jlvillal | Goneri: Thanks | 17:28 |
| openstackgerrit | Gonéri Le Bouder proposed openstack/ironic-python-agent: iscsi: wipe part table before starting the target https://review.openstack.org/284347 | 17:28 |
| Goneri | done | 17:28 |
| devananda | jroll: your api header patch needs a quick fix | 17:30 |
| jlvillal | Goneri: +2 | 17:30 |
| *** Sukhdev has quit IRC | 17:31 | |
| jroll | devananda: not seeing it, you mean in the commit message or? | 17:31 |
| jroll | devananda: or the actual header returned via the API | 17:32 |
| *** alexpilotti has joined #openstack-ironic | 17:32 | |
| devananda | the actual header | 17:32 |
| *** phuongnh has quit IRC | 17:32 | |
| *** afaranha has quit IRC | 17:32 | |
| jroll | devananda: ugh, why is that being manipulated | 17:33 |
| devananda | jroll: http://paste.openstack.org/show/LsHfZ2mn9EHGWQUOeV5C/ | 17:33 |
| jroll | devananda: anyway, you want that updated in commit message or reno or the code or all? | 17:33 |
| jroll | mmmm | 17:33 |
| jroll | I bet curl manipulates that | 17:33 |
| devananda | jroll: wait. it's just this line that's wrong: state.response.headers['OpenStack-Request-ID'] = request_id | 17:33 |
| jroll | will fix shortly | 17:33 |
| *** moshele has joined #openstack-ironic | 17:34 | |
| jroll | ? | 17:34 |
| *** lucas-afk is now known as lucasagomes | 17:34 | |
| lucasagomes | NobodyCam, :D | 17:34 |
| jroll | devananda: what's wrong with that, other than tools will munge the case? | 17:34 |
| devananda | I'll just edit the patch | 17:35 |
| devananda | jroll: it's the wrong header name | 17:35 |
| NobodyCam | lucasagomes: ahh your not going away | 17:35 |
| NobodyCam | ;) | 17:35 |
| devananda | Openstack-Baremetal-Request-ID vs Openstack-Request-ID | 17:35 |
| jroll | devananda: it should be Openstack-Request-ID | 17:36 |
| devananda | oh! | 17:36 |
| jroll | we had a discussion in the comments there | 17:36 |
| devananda | then your commit message is wrong | 17:36 |
| jroll | the commit message is wrong | 17:36 |
| jroll | yeah | 17:36 |
| jroll | I'll normalize the case too | 17:36 |
| lucasagomes | jroll, oh but the commit message needs to be updated :-/ | 17:36 |
| *** ifarkas has quit IRC | 17:36 | |
| *** alexpilotti has quit IRC | 17:36 | |
| jroll | right | 17:36 |
| * lucasagomes revoked his vote | 17:36 | |
| lucasagomes | as devananda pointed out | 17:37 |
| devananda | given the order of things, I was interpreting your latest patch as fixing things based on the discussion | 17:37 |
| devananda | cool. i'll fix since I'm already staring at it | 17:38 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Append 'Openstack-Baremetal-Request-ID' header to the response https://review.openstack.org/238008 | 17:38 |
| jroll | devananda: lucasagomes: ^ | 17:38 |
| jroll | heh | 17:38 |
| lucasagomes | jroll, first line still wrong :-( | 17:38 |
| jroll | gdi | 17:38 |
| devananda | lol | 17:38 |
| * lucasagomes changes it in gerrit | 17:38 | |
| lucasagomes | or u do it, if you have it handy :D | 17:38 |
| *** klindgren has joined #openstack-ironic | 17:38 | |
| jroll | too late | 17:38 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Append 'Openstack-Request-Id' header to the response https://review.openstack.org/238008 | 17:38 |
| lucasagomes | jroll, heh cheers, +2 | 17:39 |
| devananda | +2+A'd | 17:39 |
| jroll | thanks | 17:39 |
| klindgren | Hello - wondering if their is a feature in IPA or ironic that lets you set the device naming for nics? Since, we are trying to use configdrive, but the issue is that depending on the OS and the server the device to use for networking is totally different | 17:40 |
| klindgren | (aka I got 99 problems and systemd is one) | 17:40 |
| jroll | klindgren: udev rules in your images can help | 17:40 |
| klindgren | but this would also apply to some servers that have onboard nics being used and other servers that have 10gig addin cards that is being used | 17:41 |
| *** krtaylor has joined #openstack-ironic | 17:41 | |
| *** daemontool has joined #openstack-ironic | 17:42 | |
| *** Nisha_away has quit IRC | 17:43 | |
| gmmaha | klindgren: what OS are you using on those machines? | 17:43 |
| *** krotscheck has quit IRC | 17:43 | |
| jroll | klindgren: yeah, I'm not sure then, someone else may know :) | 17:44 |
| klindgren | Combo between cent6 and cent7 | 17:44 |
| klindgren | so systemd and non-systemd :-/ | 17:44 |
| gmmaha | aaah.. have never used cent* :( | 17:44 |
| jlvillal | klindgren: Are they coming up as eth0, eth1? Or as like p7p213? | 17:44 |
| *** pcaruana has joined #openstack-ironic | 17:44 | |
| jlvillal | biosdevname is a helpful thing | 17:44 |
| jroll | stepping away for lunch | 17:45 |
| *** derekh has quit IRC | 17:45 | |
| gmmaha | klindgren: i havent seen a feature in ironic that will do wth setting the NIC naming.. the best i found without systemd was biosdevname and with systemd (systemd itself) | 17:46 |
| lucasagomes | klindgren, hmm odd, do you have net.ifnames=0 in ur kernel cmdline? | 17:47 |
| lucasagomes | if so that would disable the consistent naming | 17:47 |
| klindgren | it depends on the os - the bigger issues is that configdrive from nova always has eth0 as configured. But the device to use might be eth0, it might be eth3 (add in 10gig card). Or in the case of systemd and biosdevname it might be em1 or em2 or p1p2 or p2p1 | 17:47 |
| *** praneshp_ has joined #openstack-ironic | 17:47 | |
| lucasagomes | klindgren, yeah :-/ it's a pain indeed | 17:48 |
| *** mdorman has joined #openstack-ironic | 17:48 | |
| klindgren | I was hoping that I could set something that indicated which nic/name should be used and the IPA could do some smarts depending on whats being done | 17:49 |
| lucasagomes | klindgren, may worth asking #tripleo how they do it, I think they use a combination of inspection + os-net-config | 17:49 |
| jroll | klindgren: check out the json network metadata, it doesn't depend on the interface name http://specs.openstack.org/openstack/nova-specs/specs/liberty/implemented/metadata-service-network-info.html | 17:49 |
| lucasagomes | klindgren, IPA won't tweak the tenants imae | 17:50 |
| lucasagomes | image | 17:50 |
| * TheJulia seconds the network metadata | 17:50 | |
| klindgren | I am not asking for it to tweek the image I am asking if it could tweek the config drive data that it creates | 17:50 |
| TheJulia | you would have to, realistically do that in nova, or in the request to create the configdrive | 17:50 |
| lucasagomes | yeah it won't be do that either because it will make the configdrive out of sync with the metadata server etc | 17:51 |
| lucasagomes | TheJulia, yeah | 17:51 |
| klindgren | but how would nova know how to set it when ironic is the thing that knows about the hardware? | 17:51 |
| TheJulia | klindgren: realistically, something in the OS needs to process the network metadata from nova | 17:52 |
| jroll | seriously, use the json metadata, cloud-init will match the macs to interfaces | 17:52 |
| klindgren | which we have cloud-init doing - but it doesn't really transform the template at all | 17:53 |
| klindgren | on kilo - so :-/ | 17:53 |
| jroll | welp | 17:53 |
| * jroll really goes to lunch now | 17:53 | |
| TheJulia | The dib simple-init element/glean will read the json and do the right thing, but you'll naturally need the metadata | 17:53 |
| mdorman | so cloud-init already supports the json network metadata? | 17:54 |
| *** trown|lunch is now known as trown | 17:54 | |
| TheJulia | mdorman: I think they are still working on support for it | 17:54 |
| mdorman | kk | 17:54 |
| lucasagomes | klindgren, right, check with #tripleo how they do that. I remember that long time ago we wanted to do something like https://review.openstack.org/#/c/145302/ | 17:54 |
| TheJulia | last I heard, it was not implemented, but I've not kept on top if it | 17:54 |
| mdorman | sure, ok. fair enough | 17:54 |
| lucasagomes | that could pass extra information from ironic to the config drive when generating it | 17:54 |
| openstackgerrit | Merged openstack/ironic: Document partition image support with agent_ilo https://review.openstack.org/295710 | 17:56 |
| lucasagomes | folks I'm going to call it a day | 17:57 |
| lucasagomes | have a great evening all! | 17:57 |
| TheJulia | goodnight lucasagomes | 17:57 |
| devananda | g'night, lucasagomes o/ | 17:57 |
| lucasagomes | see you tmrw | 17:57 |
| *** lucasagomes is now known as lucas-dinner | 17:57 | |
| *** dprince has joined #openstack-ironic | 17:57 | |
| *** d0ugal has quit IRC | 17:58 | |
| *** d0ugal has joined #openstack-ironic | 17:59 | |
| sambetts | glean definatly supports the network metadata, and I have patches in flight to support vlan and bonding with glean too, see https://review.openstack.org/#/c/289412/1/nova/virt/ironic/driver.py for a example on how to generate custom network metadata from Ironic, that patch is for supporting ports and portgroups in the neutron intergration but it certainly a start :) | 17:59 |
| *** baoli has quit IRC | 17:59 | |
| sambetts | mdorman: ^ | 17:59 |
| NobodyCam | night lucas-dinner | 17:59 |
| sambetts | o/ lucas-dinner | 18:00 |
| TheJulia | :) | 18:00 |
| *** baoli has joined #openstack-ironic | 18:00 | |
| mdorman | thanks sambetts | 18:00 |
| sambetts | mdorman: glean patches are here -> https://review.openstack.org/#/c/293648/ https://review.openstack.org/#/c/283726/ | 18:01 |
| mdorman | kk | 18:03 |
| * sambetts is heading off | 18:03 | |
| jlvillal | sambetts: Ciao! | 18:03 |
| * jlvillal heads to lunch | 18:03 | |
| sambetts | o/ jlvillal, night all | 18:03 |
| *** sambetts is now known as sambetts|afk | 18:04 | |
| JayF | jroll: need a hand with that kilo gate you were mentioning earlier? If not going to grab my upstream patchsets and go | 18:04 |
| *** cdearborn has quit IRC | 18:07 | |
| NobodyCam | night sambetts|afk | 18:07 |
| *** d0ugal has quit IRC | 18:09 | |
| *** aginwala has joined #openstack-ironic | 18:11 | |
| *** mkovacik__ has quit IRC | 18:12 | |
| *** phuongnh has joined #openstack-ironic | 18:13 | |
| jroll | JayF: nah, I'm waiting to finish out mitaka before starting on that | 18:14 |
| *** aginwala has quit IRC | 18:15 | |
| JayF | ah | 18:15 |
| JayF | need me to review anything to help with that? | 18:15 |
| *** ChubYann has joined #openstack-ironic | 18:16 | |
| jroll | JayF: anything here is up for grabs https://etherpad.openstack.org/p/ironic-mitaka-finish | 18:16 |
| openstackgerrit | Andre Aranha proposed openstack/python-oneviewclient: Extend python-oneviewclient to Uplink Set https://review.openstack.org/295981 | 18:17 |
| * jroll still afk >.> | 18:18 | |
| *** aginwala has joined #openstack-ironic | 18:18 | |
| * NobodyCam wounders who is posting for jroll as he is afk | 18:18 | |
| NobodyCam | :p | 18:18 |
| jroll | shhh.' | 18:18 |
| jroll | just checking on things between lunch and walking the dog :P | 18:19 |
| JayF | NobodyCam: he doesn't talk about it publically anymore, but he got a chip to make the IRC go directly to his brain | 18:19 |
| NobodyCam | lol... | 18:19 |
| JayF | :P | 18:19 |
| *** mbound has joined #openstack-ironic | 18:20 | |
| NobodyCam | :) | 18:20 |
| *** boris-42 has joined #openstack-ironic | 18:21 | |
| *** daemontool has quit IRC | 18:25 | |
| *** Marga_ has quit IRC | 18:36 | |
| *** jtomasek has quit IRC | 18:37 | |
| openstackgerrit | Merged openstack/bifrost: Update mitaka release notes source https://review.openstack.org/295041 | 18:39 |
| jroll | hah | 18:41 |
| *** aginwala has quit IRC | 18:43 | |
| * TheJulia thinks its time for moar coffee | 18:45 | |
| *** mbound has quit IRC | 18:46 | |
| *** athomas has quit IRC | 18:46 | |
| *** chopmann has joined #openstack-ironic | 18:50 | |
| * mgould -> home; good night, everyone! | 18:54 | |
| openstackgerrit | greghaynes proposed openstack/bifrost: Set both forms of pxe_append_params https://review.openstack.org/296007 | 18:54 |
| *** mgould has quit IRC | 18:54 | |
| *** aginwala has joined #openstack-ironic | 18:55 | |
| *** aginwala has quit IRC | 18:57 | |
| *** aginwala has joined #openstack-ironic | 18:58 | |
| *** Sukhdev has joined #openstack-ironic | 18:59 | |
| *** alexpilotti has joined #openstack-ironic | 19:01 | |
| *** Marga_ has joined #openstack-ironic | 19:02 | |
| *** mtanino has quit IRC | 19:04 | |
| *** alexpilotti has quit IRC | 19:06 | |
| *** achanda has quit IRC | 19:07 | |
| *** aswadr_ has quit IRC | 19:10 | |
| *** electrofelix has quit IRC | 19:12 | |
| *** achanda has joined #openstack-ironic | 19:16 | |
| *** ametts has quit IRC | 19:17 | |
| openstackgerrit | Merged openstack/ironic-python-agent: iscsi: wipe part table before starting the target https://review.openstack.org/284347 | 19:21 |
| *** alexpilotti has joined #openstack-ironic | 19:27 | |
| openstackgerrit | Merged openstack/ironic: Add disk_label and node_uuid for agent drivers https://review.openstack.org/295751 | 19:27 |
| openstackgerrit | Merged openstack/ironic: Fix sphinx docs build https://review.openstack.org/295868 | 19:28 |
| *** alexpilotti has quit IRC | 19:31 | |
| *** suro-patz has quit IRC | 19:32 | |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Add portgroups to support LAG interfaces - API https://review.openstack.org/206244 | 19:33 |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Update the deploy drivers with network flipping logic https://review.openstack.org/213262 | 19:33 |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: WIP: add network drivers https://review.openstack.org/285852 | 19:33 |
| *** alexpilotti has joined #openstack-ironic | 19:33 | |
| *** alexpilotti has quit IRC | 19:33 | |
| *** alexpilotti has joined #openstack-ironic | 19:33 | |
| *** harlowja_at_home has quit IRC | 19:37 | |
| *** cdearborn has joined #openstack-ironic | 19:38 | |
| *** Goneri has quit IRC | 19:47 | |
| *** piet has joined #openstack-ironic | 19:56 | |
| *** piet has quit IRC | 19:56 | |
| *** david-lyle_ has joined #openstack-ironic | 19:56 | |
| openstackgerrit | Merged openstack/ironic: Append 'Openstack-Request-Id' header to the response https://review.openstack.org/238008 | 19:57 |
| *** praneshp__ has joined #openstack-ironic | 19:57 | |
| *** chopmann has quit IRC | 19:57 | |
| *** david-lyle has quit IRC | 19:57 | |
| *** praneshp_ has quit IRC | 19:58 | |
| *** clenimar has quit IRC | 19:58 | |
| *** ekarlso- has quit IRC | 19:58 | |
| *** jlvillal has quit IRC | 19:58 | |
| *** lynxman has quit IRC | 19:58 | |
| *** lynxman has joined #openstack-ironic | 19:59 | |
| *** lynxman has joined #openstack-ironic | 19:59 | |
| *** suro-patz has joined #openstack-ironic | 20:00 | |
| *** alexpilotti has quit IRC | 20:01 | |
| *** david-lyle_ is now known as david-lyle | 20:02 | |
| jroll | does anybody have anything they want to block an ironic release on? | 20:02 |
| jroll | NobodyCam: devananda TheJulia ^ | 20:02 |
| *** dprince has quit IRC | 20:02 | |
| *** baoli has quit IRC | 20:02 | |
| NobodyCam | humm | 20:03 |
| * NobodyCam takes a quick look | 20:03 | |
| openstackgerrit | Andre Aranha proposed openstack/python-oneviewclient: Extend python-oneviewclient to Uplink Set https://review.openstack.org/295981 | 20:03 |
| *** jlvillal has joined #openstack-ironic | 20:03 | |
| devananda | jroll: wdyt of https://review.openstack.org/291829 ? | 20:05 |
| *** Sukhdev has quit IRC | 20:05 | |
| devananda | it adds a new option to the API -- so it wont be suitable to backports | 20:05 |
| JayF | looking | 20:05 |
| JayF | oooh | 20:05 |
| devananda | and it helps with the deprecation process for the bash ramdisk | 20:05 |
| NobodyCam | I'd be okay with it | 20:05 |
| NobodyCam | even thou I only have a +1 on it now | 20:06 |
| jroll | devananda: oh yeah, for some reason I thought that was in | 20:06 |
| devananda | I thought so too. just noticed it wasn't | 20:06 |
| * devananda reapplies +2 | 20:06 | |
| NobodyCam | :) | 20:07 |
| jroll | devananda: don't want to approve? | 20:07 |
| devananda | jroll: if you're good with it now, I will | 20:07 |
| jroll | devananda: I haven't reviewed it but I'm fine with the concept | 20:08 |
| devananda | jroll: it's an easy read. I'm fine waiting if you want to review it | 20:08 |
| devananda | jroll: or I can hit the button :) | 20:08 |
| jroll | devananda: go ahead, between two meetings right now :) | 20:09 |
| jroll | otherwise I can hit it in an hour | 20:09 |
| devananda | done | 20:09 |
| jroll | thanks | 20:09 |
| jroll | once that comes down I'll release, unless someone has objections | 20:09 |
| devananda | works for me | 20:09 |
| JayF | jroll: if you're only between two meetings, you're currently missing one :P | 20:09 |
| TheJulia | from my pov, I haven't seen anything | 20:09 |
| jroll | JayF: uh oh, which am I missing | 20:10 |
| JayF | jroll: open book | 20:10 |
| jroll | I'm there, plus TC meeting | 20:10 |
| *** clenimar has joined #openstack-ironic | 20:10 | |
| JayF | lol | 20:10 |
| jroll | idk what the third is | 20:10 |
| JayF | multimeetingtasking | 20:10 |
| jroll | one is ears one is eyes | 20:11 |
| jroll | :P | 20:11 |
| *** ekarlso- has joined #openstack-ironic | 20:11 | |
| * NobodyCam sees that and thinks : https://www.youtube.com/watch?v=97ECZMvbLxg | 20:13 | |
| *** aginwala has quit IRC | 20:13 | |
| *** alexpilotti has joined #openstack-ironic | 20:14 | |
| *** aginwala has joined #openstack-ironic | 20:18 | |
| *** alexpilotti has quit IRC | 20:18 | |
| *** achanda has quit IRC | 20:34 | |
| *** moshele has quit IRC | 20:37 | |
| *** ChrisAusten has joined #openstack-ironic | 20:37 | |
| *** moshele has joined #openstack-ironic | 20:38 | |
| *** aginwala has quit IRC | 20:38 | |
| *** aginwala has joined #openstack-ironic | 20:41 | |
| jlvillal | jroll: I think https://etherpad.openstack.org/p/ironic-mitaka-finish is looking pretty good. | 20:41 |
| jroll | jlvillal: yep, waiting on https://review.openstack.org/#/c/291829/ | 20:42 |
| jlvillal | There is the one IPA patch, which I gave a +1. | 20:42 |
| JayF | jlvillal: link? | 20:43 |
| JayF | jlvillal: I thought all the ipa stuff was landed | 20:43 |
| jlvillal | JayF: https://review.openstack.org/#/c/295766/ | 20:44 |
| jlvillal | It would need to be back-ported | 20:44 |
| jlvillal | jroll: I don't see https://review.openstack.org/#/c/291829/ in https://etherpad.openstack.org/p/ironic-mitaka-finish | 20:44 |
| jroll | jlvillal: see conversation above, we decided to pull it in | 20:45 |
| JayF | +2'd that ipa patch, although wouldn't hurt my feelings to get another review on that | 20:45 |
| JayF | since I'm not pro at partition images | 20:45 |
| jlvillal | Ah, okay. thanks. /me reads back-scroll | 20:45 |
| *** suro-patz has quit IRC | 20:51 | |
| *** jaypipes has quit IRC | 20:52 | |
| *** e0ne has quit IRC | 20:53 | |
| openstackgerrit | Jarrod Johnson proposed openstack/pyghmi: Provide attach remote media function https://review.openstack.org/296047 | 20:53 |
| *** spandhe has quit IRC | 20:54 | |
| openstackgerrit | Merged openstack/ironic: Agent: Out-of-band power off on deploy https://review.openstack.org/291829 | 20:56 |
| jroll | woo | 20:57 |
| *** moshele has quit IRC | 20:57 | |
| jroll | alrighty, I'm releasing the hounds | 20:57 |
| devananda | jroll: \o/ | 20:57 |
| *** alexpilotti has joined #openstack-ironic | 20:58 | |
| *** raildo is now known as raildo-afk | 20:58 | |
| jroll | https://review.openstack.org/296049 | 20:59 |
| jroll | thanks for the hard work this cycle, everyone. | 20:59 |
| *** Sukhdev has joined #openstack-ironic | 20:59 | |
| jlvillal | jroll: devananda: On https://review.openstack.org/#/c/291829/10/ironic/drivers/modules/agent_base_vendor.py Assuming it is okay that it is always doing OOB there. Or am I missing something? | 20:59 |
| *** alexpilotti has quit IRC | 21:00 | |
| jroll | jlvillal: where now? | 21:00 |
| *** alexpilotti has joined #openstack-ironic | 21:00 | |
| jlvillal | Isn't line 703 executed for all conditions? | 21:00 |
| jlvillal | Is line 703 the out of band? | 21:00 |
| jlvillal | I might just be confused. | 21:01 |
| jroll | 703 checks if it should be oob | 21:01 |
| jroll | 705 tells the agent to in-band reboot | 21:01 |
| devananda | jlvillal: default path is: if not false: allow inband reboot | 21:01 |
| jlvillal | Sorry 729 on the right | 21:01 |
| jroll | if that fails, it falls through to 729 for oob | 21:01 |
| jroll | er, 705 tells the agent to in-band *shutdown* | 21:01 |
| devananda | ooh | 21:01 |
| jroll | and 729 starts it back up | 21:02 |
| jroll | (if it succeeds) | 21:02 |
| devananda | yea, it's fine. | 21:02 |
| devananda | L706 will wait for the agent to finish powering down (or timeout) before it hits L729 and forces the reboot | 21:02 |
| jroll | it's basically, if not do_oob: try do_inband_shutdown() except: log; oob_reboot() | 21:02 |
| devananda | and L729 will just power it up if the graceful shutdown was successful | 21:03 |
| jlvillal | But it always does do OOB reboot. Correct? | 21:03 |
| devananda | jlvillal: not exactly, no | 21:03 |
| jlvillal | From L729 on right | 21:03 |
| devananda | jlvillal: it always calls that, yes. but that doesn't always do the reboot | 21:03 |
| jroll | it always *calls* oob_reboot(), which will just power it on if it is already off | 21:03 |
| devananda | ^ right | 21:03 |
| jlvillal | Oh, okay. Thanks! | 21:03 |
| jroll | 705-6 shuts it off, assuming the agent is able to do so | 21:03 |
| jroll | (which didn't change from before that patch) | 21:04 |
| jlvillal | Thanks for explaining it :) | 21:04 |
| devananda | I think the confusing bit is that manager_utils.node_power_action(task, states.REBOOT) also gracefully handles the state where the node is powered off already | 21:04 |
| jroll | right | 21:04 |
| jlvillal | That is what confused me. | 21:04 |
| devananda | jlvillal: glad we could help clear it up :) | 21:04 |
| jlvillal | :) | 21:04 |
| *** alexpilotti has quit IRC | 21:05 | |
| *** alexpilotti has joined #openstack-ironic | 21:08 | |
| *** jcoufal has quit IRC | 21:10 | |
| *** trown is now known as trown|outtypewww | 21:10 | |
| *** dims has joined #openstack-ironic | 21:11 | |
| *** alexpilotti has quit IRC | 21:13 | |
| jroll | stepping away for a while | 21:13 |
| *** Sukhdev has quit IRC | 21:19 | |
| *** suro-patz has joined #openstack-ironic | 21:27 | |
| *** keedya has quit IRC | 21:28 | |
| *** spandhe has joined #openstack-ironic | 21:29 | |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Add test to verify ironic multitenancy https://review.openstack.org/269157 | 21:34 |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Allow to build user image with DIB https://review.openstack.org/256363 | 21:34 |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Add Link-Local-Connection info to ironic port https://review.openstack.org/256365 | 21:34 |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Added operator documentation for ironic portgroups https://review.openstack.org/228496 | 21:34 |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Add portgroups to support LAG interfaces - API https://review.openstack.org/206244 | 21:34 |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: refactor ironic enroll-node code https://review.openstack.org/256364 | 21:34 |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Add configure_provision_network function https://review.openstack.org/256367 | 21:34 |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Update Ironic VM network connection https://review.openstack.org/256366 | 21:34 |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Update the deploy drivers with network flipping logic https://review.openstack.org/213262 | 21:34 |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: WIP: add network drivers https://review.openstack.org/285852 | 21:34 |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Add Ironic/Neutron integration documentation https://review.openstack.org/258596 | 21:34 |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Allow to use NETWORK_DRIVERS in devstack https://review.openstack.org/293520 | 21:34 |
| *** achanda has joined #openstack-ironic | 21:34 | |
| *** alexpilotti has joined #openstack-ironic | 21:36 | |
| *** alexpilotti has quit IRC | 21:37 | |
| *** alexpilo_ has joined #openstack-ironic | 21:37 | |
| *** achanda has quit IRC | 21:40 | |
| *** daemontool has joined #openstack-ironic | 21:41 | |
| *** alexpilo_ has quit IRC | 21:41 | |
| *** aginwala has quit IRC | 21:42 | |
| *** moshele has joined #openstack-ironic | 21:43 | |
| *** openstackgerrit has quit IRC | 21:48 | |
| *** openstackgerrit has joined #openstack-ironic | 21:49 | |
| *** e0ne has joined #openstack-ironic | 21:50 | |
| *** Sukhdev has joined #openstack-ironic | 21:54 | |
| *** aginwala has joined #openstack-ironic | 22:00 | |
| *** baoli has joined #openstack-ironic | 22:07 | |
| *** harlowja_at_home has joined #openstack-ironic | 22:08 | |
| *** mjturek1 has quit IRC | 22:10 | |
| *** baoli has quit IRC | 22:11 | |
| JayF | How receptive would folks be to a "force state" API, or something similar? | 22:19 |
| JayF | Where regardless of what Ironic thinks, an operator could "force" a machine into a given state -- I'm thinking like, forcing a deploying -> deploy failed or cleaning -> clean failed as the primary uses | 22:20 |
| JayF | sorta along the lines of the stack of operator features which are a "bad idea" in a perfect world, but as an operator I find I want because the real world is far from perfect :) | 22:20 |
| *** e0ne has quit IRC | 22:21 | |
| *** achanda has joined #openstack-ironic | 22:21 | |
| *** alexpilotti has joined #openstack-ironic | 22:22 | |
| *** moshele has quit IRC | 22:22 | |
| *** alexpilotti has quit IRC | 22:23 | |
| *** alexpilo_ has joined #openstack-ironic | 22:23 | |
| NobodyCam | JayF: sounds kinda like what the Active Node Creation spec started out as | 22:24 |
| *** alexpilo_ has quit IRC | 22:24 | |
| *** alexpilotti has joined #openstack-ironic | 22:25 | |
| JayF | I think there's a hugely different use case | 22:25 |
| JayF | in my use case, I'm thinking some kind of "outage" situation happened | 22:25 |
| NobodyCam | oh not saying that... | 22:25 |
| JayF | like a conductor being unable to reach agents | 22:25 |
| NobodyCam | just kinda reminded me of how that got started | 22:26 |
| JayF | today, that can leave nodes in weird states, including being stuck forever in "deploying" state unless you touch the db | 22:26 |
| JayF | same thing can happen in cases of agent failure | 22:26 |
| JayF | and I'm getting tired of having to write SQL to cleanup my testing environments, it's an antipattern to have to touch the db ever :( | 22:26 |
| NobodyCam | so whould this be force ANY state or more like force into *ERROR state | 22:27 |
| devananda | JayF: I agree with ^, however I have yet to see a proposal for a way to handle that which doesn't a) risk blowing up in some valid cases, or b) expose the same antipattern in an API | 22:27 |
| JayF | I mean, bluntly, I don't think that's a reasonable thing to ask: we ship software so broken it requires a crowbar to fix, but yet make operators bring their own | 22:28 |
| JayF | I appreciate the perspective of "lets make this work the Right Way(tm) in all cases so we don't need tools like this", but I don't think it's reasonable of us to continue to withhold that kind of tooling in our API | 22:29 |
| JayF | even Nova has reset-state :) | 22:29 |
| devananda | JayF: is reset-state a nova-manage command? iow, does it use the REST API, or a CLI that munges the DB for you? | 22:30 |
| JayF | that's all I'm looking for realistically, a reset-state for ironic to throw a node into error no matter what, so the operator (who might know more than Ironic about environmental failures) can tell Ironic to give us | 22:30 |
| JayF | *give up | 22:30 |
| JayF | let me see, I honestly odn't know | 22:30 |
| devananda | I'm all for an operator tool that makes this easier | 22:30 |
| devananda | but I don't want to encode it in the REST API | 22:30 |
| JayF | http://docs.openstack.org/admin-guide-cloud/support-compute.html /me glad to see it's not downstream | 22:30 |
| JayF | Oh, I'm pretty sure this hits an api endpoint | 22:31 |
| JayF | I run it from my desktop and it hits nova-api to do it | 22:31 |
| devananda | huh | 22:31 |
| alaski | it's an "admin" API command | 22:31 |
| *** Sukhdev has quit IRC | 22:31 | |
| JayF | https://github.com/openstack/nova/blob/1635086529c43edcb69966c48c41b5fca2d1b26c/nova/api/openstack/compute/admin_actions.py#L71 | 22:31 |
| alaski | if it were proposed today I'm not sure it would be accepted | 22:31 |
| devananda | JayF: yah, you're correct | 22:32 |
| JayF | I'm very glad it does exist, because it's required to run a cloud | 22:32 |
| devananda | alaski: hehe | 22:32 |
| alaski | JayF: sure, but not in the REST API | 22:32 |
| JayF | How else would it be implemented then? | 22:32 |
| alaski | nova-manage, or really I would like a separate admin endpoint | 22:32 |
| *** baoli has joined #openstack-ironic | 22:32 | |
| JayF | In a nutshell, what is nova-manage? | 22:33 |
| devananda | JayF: I believe what alaski is saying matches what i'm saying | 22:33 |
| TheJulia | NobodyCam: yeah, this is along the same exact lines of discussion ANC came up from, although it was focused on creating a node into existance that ultimately is externally deployed | 22:33 |
| *** Sukhdev has joined #openstack-ironic | 22:33 | |
| alaski | JayF: CLI interface for admins. mostly direct db interaction type of stuff | 22:33 |
| devananda | JayF: nova-manage is a CLI for wrapping db or other *nix style commands | 22:33 |
| JayF | LOL https://wiki.openstack.org/wiki/NovaManage "The nova-manage isn't properly documented, but it's going away in Folsom, so using this wiki page for quick docs on it. | 22:33 |
| *** baoli_ has joined #openstack-ironic | 22:33 | |
| devananda | things that cloud operators definitely need, but which shouldn't be exposed in a public REST API endpoint | 22:33 |
| TheJulia | nice | 22:33 |
| alaski | hah. that's not going to happen | 22:33 |
| NobodyCam | TheJulia: ya I was just saying it reminded me of how ANC came about | 22:33 |
| JayF | I would not be keen on it being some cli command that reaches out and touches my DB either | 22:33 |
| JayF | separate admin api endpoint doesn't make sense for Ironic -- as we say all the time we're already an admin-only api | 22:34 |
| devananda | JayF: "nova-manage db sync" | 22:34 |
| alaski | online db data migrations are handled via nova-manage now | 22:34 |
| NobodyCam | thou I can see value in something like force-into-error-state | 22:35 |
| devananda | it is exactly that - a CLI that touches your DB | 22:35 |
| TheJulia | NobodyCam: or force out of error state | 22:35 |
| JayF | See, I don't understand why a reset-state API endpoint is bad but that isn't | 22:35 |
| JayF | TheJulia: like nova reset-state --active | 22:35 |
| TheJulia | exactly | 22:35 |
| JayF | Might I suggest this would make a /damn good/ design summit chat at the summit? | 22:36 |
| devananda | JayF: completely agreed :) | 22:36 |
| JayF | it's an elephant-in-the-room kind of problem | 22:36 |
| devananda | yup | 22:36 |
| JayF | that ironic's operation requires hitting the db sometimes | 22:36 |
| JayF | and we all just accept it | 22:36 |
| devananda | I don't think anyone is going to argue that it's not a problem | 22:36 |
| NobodyCam | :( | 22:36 |
| JayF | and I'm tired of having to review SQL for my peers to cleanup Ironic's messes | 22:36 |
| devananda | JayF: fair enough. let's find a solution | 22:36 |
| devananda | JayF: you want to lead the session? :) | 22:37 |
| JayF | Abso-effen-lutely | 22:37 |
| JayF | well, actually | 22:37 |
| JayF | if the leader is supposed to be more of a ... moderator | 22:37 |
| JayF | probably shouldn't be me :) | 22:37 |
| NobodyCam | no take backs | 22:37 |
| devananda | JayF: haha :P | 22:37 |
| *** baoli has quit IRC | 22:37 | |
| *** Sukhdev has quit IRC | 22:37 | |
| JayF | if it's OK for the person running the session to ahve an agenda, I'm very on board | 22:37 |
| devananda | JayF: well, you should talk to jroll either way. and stick that on the topic pad | 22:37 |
| JayF | I bet jroll is as sick of this crap as I am :P and I'm looking for the topic pad right now | 22:37 |
| NobodyCam | JayF: as summit is just around the coner maybe a hacking session on fixing ??? | 22:38 |
| alaski | JayF: I'm not arguing against a way to do what you want. Just offering my perspective which is that embedding reset-state in the API makes it a permanent fixture, and adds an expectation that it should be used. Having it as a low level tool to accomplish it sets the right expectations about how often it should be used | 22:38 |
| devananda | sessions usually need an empassioned presenter, a moderator, and a note-taker. sometimes more than one of each :) | 22:38 |
| JayF | NobodyCam: +.5, like here's the deal: we'll always be chasing the next bug | 22:38 |
| JayF | NobodyCam: we have to give operators the tool to unwind all the bugs we haven't written yet | 22:38 |
| devananda | alaski: ++ | 22:38 |
| JayF | alaski: I will agree that I think reset-state can be ... overused | 22:38 |
| devananda | NobodyCam: -- on a hacking session to fix it because we haven't even agreed on HOW to fix it | 22:39 |
| JayF | alaski: but given our API has node-delete, with no warning, that's arguably much more destructive, I find it hard to get upset about :) | 22:39 |
| TheJulia | alaski: agreed, although it is also unreasonable to expect all admins to have direct database access to do their jobs, it can become a separation of duties issue in some orgs very quickly | 22:39 |
| devananda | JayF: til; we need policy support in our API | 22:39 |
| devananda | NobodyCam: now if you wanted to hack out policy support, I'd be on board with that | 22:40 |
| devananda | also, apparently, it won't be hard for us to add | 22:40 |
| devananda | I just haven't had th etime | 22:40 |
| TheJulia | the framework is there | 22:40 |
| alaski | JayF: fair, but having one mess doesn't mean others should be added :) | 22:40 |
| JayF | devananda: ++ with policy support too, heh | 22:40 |
| TheJulia | just takes time to define and punch into place | 22:40 |
| devananda | TheJulia: right | 22:40 |
| JayF | alaski: I'm more making the point there's a big puddle on the floor and everyone is arguing over the color of the mop :P | 22:40 |
| JayF | I would LOVE to hack on policy support on Friday morning at the summit | 22:41 |
| alaski | TheJulia: that's fair. I am also amenable to a separate admin endpoint, though I understand that's an odd distinction in the ironic case | 22:41 |
| JayF | is it a code-date? | 22:41 |
| JayF | :) | 22:41 |
| TheJulia | JayF: shopvac | 22:41 |
| * NobodyCam likes the blue mop | 22:41 | |
| TheJulia | lol | 22:41 |
| JayF | TheJulia: ironic shopvac $UUID # magically cleans up whatever mess has been made of the node | 22:41 |
| NobodyCam | lol | 22:41 |
| NobodyCam | +++ | 22:41 |
| JayF | TheJulia: Backend: just files a ticket to have ops touch the db | 22:41 |
| TheJulia | lol | 22:41 |
| JayF | lol | 22:41 |
| NobodyCam | but we should call it a wetDryVac | 22:42 |
| TheJulia | JayF: 3 months later, ops replies, can't find instance | 22:42 |
| * NobodyCam *ducks* | 22:42 | |
| TheJulia | NobodyCam: I think we need to go yarn shopping in Austin.... Just saying.. | 22:43 |
| *** achanda has quit IRC | 22:43 | |
| NobodyCam | :) | 22:43 |
| *** ayoung has quit IRC | 22:43 | |
| JayF | https://etherpad.openstack.org/p/mitaka-ironic-design-summit-ideas topic added as #9 | 22:44 |
| *** rbudden has quit IRC | 22:44 | |
| *** Sukhdev has joined #openstack-ironic | 22:44 | |
| *** absubram has quit IRC | 22:47 | |
| *** aginwala has quit IRC | 22:48 | |
| * jlvillal thought JayF added section on Yarn shopping :) | 22:53 | |
| NobodyCam | lol | 22:53 |
| TheJulia | no, the yarn is to toss at NobodyCam :) | 22:53 |
| jlvillal | :) | 22:53 |
| NobodyCam | lol | 22:53 |
| jlvillal | JayF: On your topic. Seems like a superset of what TheJulia was/is doing with regards of being able to move manageable states to provisioned. | 22:56 |
| jlvillal | If I'm remembering things correctly. | 22:56 |
| TheJulia | jlvillal: What JayF is bringing up, is exactly how it got started, just the primary case we brought forth was creating nodes in inventory that are "active" | 22:57 |
| jlvillal | TheJulia: Thanks. Agreed it would be nice to just move things with some super command. Instead of mysql commands :) | 22:58 |
| JayF | My angle is completely centered around this user story: "As an operator of Ironic, I should never need to run manual SQL queries" | 22:58 |
| TheJulia | JayF: ++ | 22:59 |
| JayF | and today we have at least 2 states I'm aware of where that can happen | 22:59 |
| JayF | and I suspect it's a similar pattern for some of our async things | 22:59 |
| *** baoli_ has quit IRC | 23:00 | |
| *** baoli has joined #openstack-ironic | 23:01 | |
| *** aginwala has joined #openstack-ironic | 23:01 | |
| *** jaybeale has quit IRC | 23:02 | |
| JayF | https://review.openstack.org/#/c/293781/1 has 6x +1s if someone wants to give it a core review; one line change | 23:05 |
| cinerama | i was really excited about the yarn shopping :( :) | 23:08 |
| NobodyCam | lol | 23:09 |
| *** cdearborn has quit IRC | 23:11 | |
| *** dims_ has joined #openstack-ironic | 23:12 | |
| *** rbudden has joined #openstack-ironic | 23:13 | |
| *** mdorman has quit IRC | 23:13 | |
| *** causten_ has joined #openstack-ironic | 23:14 | |
| *** dims has quit IRC | 23:15 | |
| *** dims has joined #openstack-ironic | 23:15 | |
| TheJulia | well, 7 +1s now | 23:15 |
| *** baoli has quit IRC | 23:16 | |
| jroll | JayF: devananda: I'm always happy to moderate a session that is led by someone with strong opinions :) | 23:16 |
| *** ChrisAusten has quit IRC | 23:17 | |
| jroll | +2 for 293781 | 23:17 |
| NobodyCam | and now 1 +2 :) | 23:17 |
| *** izaakk has quit IRC | 23:17 | |
| jroll | pfft | 23:17 |
| * jroll +A | 23:17 | |
| NobodyCam | lol | 23:17 |
| jroll | :) | 23:18 |
| *** dims_ has quit IRC | 23:18 | |
| *** causten_ has quit IRC | 23:19 | |
| TheJulia | goodnight folks | 23:19 |
| jroll | \o | 23:19 |
| *** dims_ has joined #openstack-ironic | 23:19 | |
| *** achanda has joined #openstack-ironic | 23:22 | |
| *** dims has quit IRC | 23:22 | |
| *** dims has joined #openstack-ironic | 23:23 | |
| NobodyCam | have a good night TheJulia | 23:23 |
| *** jaybeale has joined #openstack-ironic | 23:24 | |
| *** dims_ has quit IRC | 23:25 | |
| gmmaha | Hi, could someone shed some light on possible pitfalls using bifrost along with ILO? | 23:28 |
| *** manjeets has joined #openstack-ironic | 23:29 | |
| devananda | gmmaha: are you already familiar with using bifrost and ipmi? | 23:29 |
| gmmaha | i am trying to use bifrost to privision some machines that are managed by ILO and the machine starts download the coreos image and jsut hangs.. nothing beyond that | 23:29 |
| gmmaha | devananda: yup.. ipmi is the one that is used as the driver agent | 23:29 |
| gmmaha | and the target machine reboots fine, but hangs with the kernel download | 23:30 |
| devananda | gmmaha: I'm having trouble parsing "is used as the driver agent" | 23:30 |
| gmmaha | and nothing beyond tath | 23:30 |
| *** rbudden has quit IRC | 23:30 | |
| gmmaha | devananda: sorry what i meant is, i am using ipmi as the method to talk to the bare metal.. in the json file that is used to enroll machines in bifrost | 23:30 |
| devananda | gmmaha: what driver is ironic using to manage the machines? "agent_ipmitool" ? | 23:30 |
| * gmmaha goes to find the exact string | 23:30 | |
| gmmaha | devananda: yes.. | 23:31 |
| devananda | ok. then this should have nothing to do with iLO | 23:31 |
| devananda | gmmaha: you might want to change the serial console redirection that the ramdisk is performing, to help with troubleshooting | 23:31 |
| gmmaha | devananda: ok. will try that out.. thanks | 23:32 |
| devananda | gmmaha: for instance, I have this option in my bifrost configs: extra_kernel_options: "nofb nomodeset vga=normal console=ttyS2,115200n8" | 23:33 |
| devananda | the exact option will vary between hardware vendors | 23:33 |
| *** praneshp__ has quit IRC | 23:34 | |
| gmmaha | devananda: aaah Ok.. let me set it such and see what happens | 23:34 |
| gmmaha | thanks | 23:34 |
| *** Sukhdev has quit IRC | 23:36 | |
| *** smoriya_ has joined #openstack-ironic | 23:48 | |
| *** suro-patz has quit IRC | 23:54 | |
| *** baoli has joined #openstack-ironic | 23:54 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!